# About Capstra Labs | Vocapable > Vocapable is a product of Capstra Labs, LLC, a California company. Who builds the agent, what runs on hardware we own, and why the whole rate card is public. URL: https://www.vocapable.com/about The company # Who builds this, and why the price is public Vocapable is a product of Capstra Labs, LLC: AI voice agents that place and answer business phone calls on the customer’s own Twilio account and their own numbers. Every call returns a record, and the compliance controls behind it cannot be switched off. ## The plain facts - Vocapable is a product of Capstra Labs, LLC. - Capstra Labs, LLC is a California limited liability company doing business as Vocapable. - Vocapable sells AI voice agents for US business phone calls. - Vocapable customers place their calls on their own Twilio account, at the carrier rates they already negotiated. - Every Vocapable call attempt returns an evidence file, and every conversation returns a transcript, a disposition and a structured outcome. Agreements are governed by the laws of the State of California. Written contact issupport@capstralabs.com, and the public code organisation ishttps://github.com/CapstraLabs. Vocapable runs on Twilio; no partnership or endorsement is implied. ## What we build The agent that holds the conversation. The campaign engine that keeps working your list after the team goes home. A CRM built for phone outreach rather than a deal pipeline with a dialer bolted on. A/v1 API for product resources, with signup, onboarding, and first-key setup in the browser. And an evidence file for every call attempt, so you can show your work later. The speech recognition, the voice and the language model all run on hardware we operate. No hosted-voice API sits in the middle taking its cut, which is why we can print the whole rate card on a public page instead of quoting you per deal, and why the per-minute number on it is as small as it is. The thesis underneath all of it: compliance is the product, not the paperwork around it. Four controls run in code on every dial, on every plan, at no charge, and nothing can turn them off. Not a plan upgrade, not an API field, not us on a support call. That is what makes your call log hold up when somebody asks to see it. ## You can evaluate us without talking to us The four things a buyer usually has to extract from a sales call are already published. Read them in an afternoon, decide, and then write to us with a real question. ### The whole rate card Every figure we bill is on the pricing page: the monthly base of each plan, the per-call-attempt fee, the per-minute AI rate on both commitment terms, and the volume ladder with the floor it stops at. The calculator on that page prices your campaign rather than a sample one. ### The security posture, control by control The Trust Center describes the security and data posture in plain prose, control by control. Most security questionnaires are already answered there before you send one. ### Every agreement, in full The master agreement and each document it pulls in are published as pages you can read now. Your own lawyer can see exactly what you would be signing. ### A sample evidence file A sample file for one call attempt, rendered the way the platform records it: which policy values bound the call, what the agent said, and what came back. This is the artifact you hand to somebody who asks how a call happened. The security and data posture, control by control, is on theTrust Center, and every agreement is published in full at/legal. ## Who is behind it Joseph Opanel Founder Joseph Opanel founded Capstra Labs, LLC in 2026. ## Read the product, not the pitch The sandbox is free and needs no card. Build an agent from your own pricing page and docs, run your script down the whole dial path with every gate live, and read the file it hands back. Everything else we publish is onthe shelf. --- # Product resources, available by API | Vocapable > Product resources use the same /v1 API; signup, onboarding, and first-key setup use browser flows. Scoped keys, mock test telephony, signed webhooks. URL: https://www.vocapable.com/api Headless # The dashboard is a client Product resources use the same /v1 API your integration calls. Create agents, run campaigns, receive structured outcomes, and bring inbound caller requests into your own workflow. Signup, onboarding, and first-key setup run in the browser. Once you have a key, your team can automate the resources and scopes it authorizes. ## The loop, end to end Launch a call with a question that matters today. The agent asks it, records the answer, and, for an enabled event, queues a signed result to your configured endpoint. Test keys route telephony to a mock provider, so your team can safely iterate on the integration without a carrier call. ## Built for production integrations ### Results come to you One signed webhook endpoint per tenant. Send call outcomes, dispositions, and structured survey or shift results into your stack, inspect delivery attempts, rotate the secret, and redrive an exhausted delivery yourself. ### Start from a working workflow Instantiate a template for lead qualification, staffing availability, shift reminders, candidate follow-up, customer feedback, or sales outreach. Make its questions, objectives, knowledge, and voice your own, then publish the version you approved. ### Turn inbound calls into work Receptionist tools can take a message, open a support request, schedule a callback, or update a known caller’s lead status. The native Inbox keeps the resulting work visible to your team and can deliver approved actions to your endpoint. ### Errors you can branch on Every non-2xx is an RFC 9457 problem whose type URI resolves to a public registry page, one page per code, hundreds of codes, kept in lockstep with the source by a build gate. Your error handling reads like a switch statement, not a regex. ### A compliance floor with no override AI disclosure plays before the model speaks. Opt-outs are honored by a detector outside the model. Every dial passes the dialability gate, and a check that cannot complete blocks the call. No API field, plan tier, or support ticket switches any of it off, which is precisely what makes the rest of the API safe to automate. ## One surface, and we can show our work The control plane assembles a single list of consumer routers and mounts it at/v1: one list, and it is the same list your key calls. The reference is generated from the same OpenAPI document the dashboard's own client is typed against, so it cannot drift from what actually answers. --- # Compliance literacy for AI calling | Vocapable > Writing on the rules that govern AI calling, what a call actually costs, and what a per-call record has to prove. Sourced, dated, and re-verifiable. URL: https://www.vocapable.com/blog Writing # The rules, the money, and the record Two kinds of writing live here. Playbooks for the people who have to buy this and explain it internally, and plain readings of the rules that govern calling — which is the half worth your time, because it is where a vendor either knows the material or is quoting somebody who did not. Every external fact carries the source it came from and the date we last read it. Statutes get amended and rules get vacated, and a citation with no date is a claim with no expiry. - Compliance August 18, 2026 Cross-channel revocation and the 2027 rule One stop request, honored everywhere it covers. Ahead of 2027. Vocapable writes one revocation, honored across every channel it covers, ahead of the 47 C.F.R. § 64.1200(a)(10) cross-channel mandate effective 2027-01-31. Read it - Sales playbook August 18, 2026 What a per-minute quote leaves out Every line on a Vocapable invoice, and where the rest stays. Every line a Vocapable invoice carries, where your carrier minutes stay, and the cost categories to ask any AI calling vendor about before you sign anything. Read it - Compliance August 18, 2026 What is in a per-call evidence file A waived stage is a signature. It is never a silent pass. The anatomy of one per-call evidence file: what bound the dial, which checks ran, and why a waived stage never reads as a check that found nothing. Read it ## The pages behind the writing Nothing above is an announcement. Every mechanism these posts describe is documented on a product page, and every price is on the rate card rather than retyped into prose where it can go stale. ### The compliance engine The controls with no off state, the single gate every dial passes, the consent taxonomy, and the one revocation write path. ### The per-call evidence file The anatomy of one record, with a synthetic sample rendered key for key beside the page. ### The published rate card Every rate we bill, in full, with an estimator that computes a campaign from the same arithmetic the bill uses. ## Read the product, then run it yourself The sandbox is free and needs no carrier account. The compliance floor runs there exactly as it runs in production. --- # Cross-channel revocation and the 2027 rule | Vocapable > Vocapable writes one revocation, honored across every channel it covers, ahead of the 47 C.F.R. § 64.1200(a)(10) cross-channel mandate effective 2027-01-31. URL: https://www.vocapable.com/blog/cross-channel-revocation-and-the-2027-rule # Cross-channel revocation and the 2027 rule One stop request, honored everywhere it covers. Ahead of 2027. Published by Vocapable, a product of Capstra Labs, LLC. A revocation is one fact about a person, not a per-channel preference. Somebody who says stop calling me has not asked to keep hearing from you somewhere else, and a system that stores that fact once per channel has already decided otherwise on their behalf. ## The rule, stated exactly 47 C.F.R. § 64.1200(a)(10) requires a revocation received through any medium to be honored across every channel the caller uses. The cross-channel portion of that requirement becomes mandatory on 2027-01-31, after two delays. Read it once and the design consequence is hard to miss: the regulation does not care which door the request arrived through, so a stack that answers per door is going to be wrong at one of them. ## Two numbers, and they are not the same number This is where writing on the rule most often goes wrong, and the error is expensive because it travels from a blog post into a contract. The rule allows up to ten business days to honor a revocation once it is received. That is a statutory bar: a ceiling on lateness, measured against a record after the fact. Vocapable’s suppression lands in the same transaction that records the revocation. That is a property of a write path, and it is not a restatement of the rule. Those are two different numbers about two different things. Merged one way, a ceiling on lateness reads as a demand for immediacy the regulation never makes. Merged the other way, a property of our write path gets restated in the statute’s units, which understates the software and gives it room to drift toward a deadline it has no need of. Your Telecom Compliance Addendum §4.1 covers the revocations that reach you off the platform: the person who tells your rep to stop, in an email, on a form we never see. Everything that arrives through the platform lands in the write described below, which is the half of the problem you no longer have to staff. ## Why a channel column is where this breaks The common shape is a preferences table with a row per person per channel, written by whichever subsystem happened to take the request. Two failure modes follow from that shape, and both are ordinary engineering rather than negligence. The write that lands in one subsystem does not reach the other, because nothing in the schema says it has to. And the read that decides whether to dial asks a different question from the read that decides anything else, so the two can disagree for as long as it takes somebody to notice. Neither is visible until a person who revoked gets contacted again — which is precisely the moment it stops being an internal defect and becomes a record somebody else gets to read. ## One write path, and no channel column Vocapable holds one suppression row and it carries no channel column. There is nothing to propagate between, because there is no second place for the fact to live, and both readers ask the one row the same question. Every entry point converges on the same transaction: a mid-call opt-out, a keypress, an inbound stop request, and your own API call. One write, whose statement order carries the guarantee. The internal do-not-call entry, the event, the withdrawal of the consent standing behind that number, and the dial-time fast path land together, or none of them land. A revocation whose covered set cannot be resolved covers everything. It over-suppresses rather than under-suppresses, which is the only direction it is allowed to err in. And it cannot be given an expiry. The endpoint refuses one rather than honoring it, because two of those writes are irreversible — the consent ledger is append-only and the dial-time flag is never re-derived — so a time-limited revocation would destroy the consent evidence permanently and then hand the number back as dialable. ## Why we ship it before the date The cross-channel requirement is mandatory on 2027-01-31. We propagate today. That is not eagerness. Gating the behavior on a date would mean shipping the non-propagating version first and then trusting somebody to remember. The date is recorded as evidence and is deliberately not branched on: there is no code path that behaves one way before it and another way after. ## What lands in the record A revocation is not only a state change; it is an entry in the evidence file for every attempt that follows it. The next dial for that number carries the internal-suppression result read live at dial time rather than from cache, and where the dial is refused, the reason on the attempt names it. The per-call anatomy is walked in full on the evidence file, and the engine that writes it is described on the compliance page. The duty split is in yourTelecom Compliance Addendum, published on this site. ## What to ask, of us or of anyone - Where does a revocation get written, and how many places store it? Any answer above one is the answer. - Is there a channel column on that row? If there is, ask what keeps the copies in agreement, and how anyone would find out that they stopped being. - Can a revocation be given an expiry? If it can, ask what happens to the consent record it withdrew. - What does the system do when it cannot resolve which channels a revocation covers? Over-suppressing is the only safe direction, and it should be the documented one. - Which of the vendor’s stated timings is the statute and which is the software? A vendor who quotes one number has merged two facts, and you cannot tell from the outside which way. What this does is make what your system did with a person’s request reconstructible by somebody who was not there, which is the part that is still true on the day you are asked to prove it. ## Sources, and the date each was read Statutes get amended and rules get vacated. A citation carrying no read-date is a claim with no expiry, so each one below records when we last confirmed it says what we say it says. - 47 CFR § 64.1200 — Delivery restrictions read 2026-08-18 The page behind this ## Compliance enforced in code Four controls no plan disables, and a file for every call. --- # What a per-minute quote leaves out | Vocapable > Every line a Vocapable invoice carries, where your carrier minutes stay, and the cost categories to ask any AI calling vendor about before you sign anything. URL: https://www.vocapable.com/blog/what-a-per-minute-quote-leaves-out # What a per-minute quote leaves out Every line on a Vocapable invoice, and where the rest stays. Published by Vocapable, a product of Capstra Labs, LLC. A per-minute figure is the easiest thing in this category to compare and one of the hardest to compare correctly, because two vendors quoting the same number can be pricing entirely different work. This is what sits behind ours, stated in the order a buyer actually needs it: what our invoice carries, what it does not, and the questions that make any vendor’s quote comparable to any other. No figure appears below. Every rate is published on the pricing page, and a number retyped into prose is a number that goes stale somewhere nobody is looking. ## What our invoice carries A Vocapable invoice has a short list of line kinds, and the rate behind each one is published rather than quoted per deal: - The monthly plan price for your workload class. - Additional concurrent sessions, at a flat monthly price, in any quantity. - A per-call-attempt fee, on the outbound plans. - AI usage per minute on the calls your account places, at the month-to-month rate or the lower committed rate, stepped down by the published volume discount as monthly spend rises. Inbound answering is priced on a different shape, and the difference is not a discount. A receptionist answers calls rather than placing them, so there is no call attempt to charge a fee against, and the per-minute AI-usage charge applies to the calls an account places rather than the ones it answers. A receptionist invoice is its monthly plan price plus its additional simultaneous connections, and that is the whole of it. ## Where the rest of your money stays - Your carrier minutes stay on your carrier’s bill, at the rates you negotiated. Every production call leaves a Twilio subaccount you own, and your carrier invoices that service to you directly, so there is no carrier line on ours to reconcile. - Your numbers stay yours. You procure and hold them on your own account, so the caller-ID reputation a year of calling earns is an asset you keep. - Going live carries no implementation or onboarding charge, on any plan. - The compliance floor is free on every plan. The controls that have no off state run at no cost, and the local floor stages are excluded from the metered lookups by construction rather than by a billing rule somebody has to remember to apply. ## The correction most people get backwards Here is the subtle part, and it is worth its own section. Our per-minute rate prices AI usage. It does not price carriage. Those are different goods that happen to be measured in the same unit, and sharing a unit is exactly why the comparison goes wrong. So a minute on a Vocapable call is metered and it is priced — as AI usage, never as carriage. What is absent from the invoice is the carrier line: the minutes your carrier moved, the number rentals, the surcharges. Those sit on your own carrier bill, at the rates you negotiated, because that account was never ours to hold. Stated precisely, because the imprecise version is a claim we would have to walk back: a Vocapable invoice carries no carrier line. That single fact is why one per-minute figure and another are not necessarily the same quantity, and why comparing two of them without first asking which goods each one covers is arithmetic performed on two different units. The mechanics of who holds the account are onthe telephony page. ## The cost categories to ask any vendor about The useful version of this section is a list of questions, and every one of them works on us too. Take it into any vendor conversation you are already having. - Is the carrier leg inside the quoted minute, or billed separately, and by whom? A rate that bundles carriage is not comparable to one that does not, and the gap is not small. - Whose name is on the carrier account? It decides who holds the registrations, who answers an industry traceback, and who keeps the caller-ID reputation when the software contract ends. - Who pays for the numbers, monthly, and what happens to them if you leave? A number you do not hold is a switching cost that appears on no quote. - What does an attempt that nobody answers cost? Unanswered dials are most of the volume in outbound calling, and a per-connected-minute quote says nothing at all about them. - Is concurrency included, capped, or metered — and what happens at the ceiling? A queue, a refusal and a charge are three very different products. - Is any compliance work metered per lookup, and can any of it be switched off to save money? A control with a price attached has an off state by definition, and the off state is the thing you are actually buying. - Is the rate card published, or assembled per deal? Ask what the renewal rate is, in writing, before you sign the first one. - What is the term, is it cancellable, and what is the refund posture? Ours is month to month, or a committed term that is non-cancellable with all sales final under a signed non-refund waiver — and we would rather you read that here than discover it in an order form. - What does it cost to get your data out, and in what format? Export should be a resource on an API, not a professional-services engagement. - What is the implementation or onboarding charge? Ask for it as a number, not as a range. ## Why we publish the card at all Speech recognition, speech synthesis and language-model serving run on infrastructure this company operates rather than on a rented hosted-voice API. That is an unusual choice in this category, and it is the reason the rate card can be published in full instead of assembled per deal. A published card is also the only kind of price a buyer can hold us to later. So run your own numbers rather than ours: the estimator onthe pricing page computes a campaign from the same arithmetic the bill uses, and the terms it prices are in theMaster Services Agreement, published on this site. The page behind this ## Pricing Published plans and usage rates. The sandbox is free to start. --- # What is in a per-call evidence file | Vocapable > The anatomy of one per-call evidence file: what bound the dial, which checks ran, and why a waived stage never reads as a check that found nothing. URL: https://www.vocapable.com/blog/what-is-in-a-per-call-evidence-file # What is in a per-call evidence file A waived stage is a signature. It is never a silent pass. Published by Vocapable, a product of Capstra Labs, LLC. A demand letter asks one question: what happened on this call, on this date, to this number. Every field described below exists to answer it without anybody’s memory being involved, and without access to the live system. ## One attempt, one file Every attempt the platform places or refuses produces exactly one policy snapshot. A call with no attempt behind it carries its snapshot on the call instead. One home per call, never both and never neither — which reads like an implementation detail right up until you are the person trying to prove a negative years later. ## What the file records - The identifiers, and the runtime and pipeline versions pinned, so the call is reproducible. - The disclosure: its exact compiled text, and playback confirmed by media-stream offset rather than by an intention to play it. - Normalization: the number as validated to E.164, rejected as a short code, or deduped against the run. - Internal suppression: your own do-not-call list applied when the list was prepared, and read live again at dial time, never from cache. - Consent: the record relied on, with its class, its scope, its source and a content hash over the proof — or the exemption class instead. - The values that bound the dial: the calling window evaluated in the contact’s local time, the attempt cap, the consent standard, and the recording-consent class. - Where each of those values came from, because a value we applied must never read as a value we vouched for. - The elections: every supplier-backed stage you run upstream, rendered as waived with its attestation reference and the person who signed it. - The outcome: the disposition, the opt-out and transfer events, and the suppression write recorded as written, failed, or not applicable. ## Waived is a signature, not a silence This is the rule the whole artifact turns on, and it is the one we would refuse to compromise if a customer asked us to. You run whatever screening arrangement you have chosen, upstream, on your own list, and you sign one attestation naming it, its signer and its date. Every stage that attestation covers appears in the file as an election rather than as a result. The stage name, the word waived, the attestation reference, and the signer’s name and role sit together on the row. A reader who was not there can see exactly who decided the stage would not run here, and on which document they decided it. A check that did not run must never read as a check that found nothing. Those are opposite facts, and a file that renders them identically has destroyed the only information the reviewer came for. ## A missing check with no election is a bug The mirror of that rule matters more than it sounds. An account carrying no attestation for the stages it runs upstream screens its contacts undialable, with the missing document named, so the dial waits for a signature. That is the difference between a system that fails closed and one that fails quiet. A gap in the record is not a pass and is never rendered as one. ## Three states a reader has to tell apart Checked, and clear. Checked, and suppressed, with the reason on the row. Not performed here, under a named election with a signature attached. A record that cannot tell those three apart is not evidence. It is a log. ## What the file is worth on the day you need it You and your counsel set the standard your calling has to meet. From there the platform does the work: it applies the values you set, blocks the calls that must not happen, and writes down what bound every one of the rest. That is what turns a demand letter from a fire drill into a lookup. The answer was written on the day of the call, by the system that placed it, instead of being reconstructed months later by whoever still remembers the campaign. ## How it gets out The snapshot rides as a column on the attempt row and leaves through the exports API as CSV, like every other resource on the platform. The audit tables behind it are append-only: no database role holds UPDATE or DELETE on them, so a correction is a new row referencing the one it corrects rather than an edit to history. Retention is set by the platform rather than by the tenant, and compliance evidence is excluded from tenant-initiated deletion, because the file is the platform’s own defense record as much as yours. A synthetic sample, rendered key for key, is published onthe evidence page. The engine that writes it — the controls with no off state, the single dialability gate, and the consent taxonomy behind the record — is onthe compliance page, and the responsibility split is stated in yourTelecom Compliance Addendum. The page behind this ## The per-call evidence file Every call leaves a file you can hand to a reviewer later. --- # Changelog: what shipped, and when | Vocapable > Every capability we shipped, dated and newest first. Each entry says what changed, who it is for, and what your business gets from it. URL: https://www.vocapable.com/changelog Company · Changelog # What shipped, and when Each entry says what changed, who it is for, and what your business gets from it. Everything on this page has shipped and is in the product now. A capability lands here on the day it ships, so the newest line is the newest thing the product does, and the date beside it is the date that became true. ## August 2026 ### A receptionist connection is prepaid, then reserved The reserved line is paid for before the slot is withheld. Adding a concurrent inbound connection now charges a full month at purchase, whatever day of the billing period it falls on, and the connection is not reserved until that charge succeeds. Why it works that way. A connection is not a feature flag. It is a slot withheld from every other claim on the fleet from the instant it is granted, busy or idle, so granting one ahead of payment gives away the scarce thing the price exists to ration. The charge is a full month rather than a fraction for the same reason: the reserve itself is not scaled to the days left in the period, so a fraction would price less than the platform actually gave up. What you are told before you agree. The confirmation step states the exact amount and the exact date your subscription next renews, because a connection bought shortly before a renewal pays its prepaid month and then renews. That is disclosed rather than smoothed away. Your existing lines are untouched. A connection you already hold keeps the dunning grace it has always had, so a receptionist you already pay for keeps answering while a payment retries. And if a new purchase fails on your own card, your connection count, your billing items and your line’s admission all stay exactly as they were. ### The AI receptionist, and its answering-policy engine Inbound calls answered on your own Twilio, in your own order. Vocapable now answers inbound calls as well as placing them. The receptionist runs on your own verified Twilio subaccount and the number you already publish, and it is sold on its own or attached to an outbound subscription. The whole policy engine ships, not a first slice. One answering policy per number, carrying an ordered chain for your open hours and another for everything else, built from the step types you compose: a keypress menu, ringing your own people, the AI agent, voicemail, a spoken message, and a forward to a number you configured. Open hours and holiday dates are yours, in your own timezone. The document is read, written and removed through /v1 under the telephony scope you already hold. Who it is for. A business that publishes a number and wants it answered in the order it decided on. Only the AI step claims a reserved inbound connection — menus, rings, forwards and spoken messages claim none — which is why a single connection can carry a far larger flow than the concurrency it bought. Your own campaigns can never take your receptionist’s line. Buy it on its own, or attach it. A receptionist-only account is priced for answering; outbound dialing comes with an outbound plan, and the two attach for the same monthly price. One piece of arithmetic admits every call on the account, inbound and outbound, and it subtracts your reserved inbound connections first, so there is only ever one gate to reason about. When every connection you bought is busy, the caller advances to the next step you wrote down, usually your own voicemail. The receptionist page carries the shape in full, and pricing carries what it costs. ### Cold outreach, admitted under an executed waiver Purchased and rented lists, admitted under a waiver you sign. Run the list you already own. Contacts you purchased, rented, or had appended now call under the Cold Outreach & Data Provenance Waiver you execute, which is the same instrument that admits an artificial-voice voicemail and a call to someone from whom you hold no prior consent. Who it is for. Accounts whose program is genuinely theirs — their data under their own supplier agreement, their carrier, their targeting, their script. The screen admits a closed list of conditions under the waiver, each rides out on the verdict by name, and electing one without the executed waiver refuses rather than proceeding. Production activation re-verifies the document at the served bytes, so the checkbox on the screening form is an intent to sign and was never an agreement. What the waiver covers. Contact data you hold under your own supplier agreement: purchased, rented, appended. The bar is provenance you can name, and an applicant who cannot name theirs reaches a human review before anything dials. The four controls enforced in code run on this traffic exactly as they run on every other call, so admitting cold outreach costs you none of the floor you already have. A signature records who carries responsibility for the program. Whether the call was one you were entitled to place is yours to determine with your own advisors. The Trust Center states the same division, and the waiver itself is published in the legal set so you can read it before you sign it. ### One published rate card, priced on AI usage One base per plan, one attempt fee, one AI-usage minute rate. The card a buyer can price a campaign against is now a monthly base per plan with included concurrent sessions, a flat add-on for each further session, a per-attempt call fee, and a per-minute rate for AI usage that steps down as total monthly spend grows. Every figure is published on the pricing page rather than quoted on request, and each one is read from the same constants the rating engine bills from. Who it is for. Anyone comparing us against a vendor whose price is a conversation. Month-to-month and the signed committed term share a base fee and differ only in the per-minute rate; the annual term was retired as a product and its plan ids were deleted rather than repriced, so no subscription can inherit a commercial promise nobody made. What the per-minute rate prices. It prices the AI session — the GPU, the speech and the language work — and not the carrier leg. That is why a completion on your own Twilio account emits an AI-usage minute exactly as any other completion does, and why a Vocapable invoice carries no carrier line: your own carrier bills the telephony directly, at the rates you negotiated. Telephony ownership is the page that owns that division. A refusal costs you nothing. An attempt the compliance gate stopped before any dial existed emits no charge at all, so the floor you are protected by never reaches your invoice. ### Developer documentation Quickstart, guides, and a reference generated from the spec. A developer documentation site: a quickstart, guides for headless mode, authentication and scopes, webhooks, conventions, test mode, and billing and meters, plus a per-operation reference for the whole /v1 surface. Who it is for. The person deciding whether this gets integrated at all. Every dashboard action is a /v1 call, so the reference is the product rather than an integration surface bolted onto one — the API page makes that argument, and the docs are where you check it. How it is built, and why that matters. The reference is generated from the same published OpenAPI document the dashboard’s own client is typed against, so every route in it is a route that answers, and it moves the day the API does. The error registry is generated the same way: every problem code the API returns resolves at the exact path its type URI points at. ### Your screening arrangement, recorded on every call Sign once, and every call names your arrangement and its signer. Keep the screening arrangement you already run, and the pricing you negotiated for it. Before your first production dial you sign one attestation naming that arrangement, its signer and its date. Its reference and its signer are written onto your compliance profile from then on, and each stage that attestation covers rides into every call’s evidence file as a named election carrying the document and the officer behind it. Six months later, when somebody asks who screened this number, the answer is on the call rather than in an email thread. Who it is for. Every production account. You keep the supplier you already hold, you never pay twice for a lookup you already buy, and your posture is never locked to a vendor somebody else picked for you. The record cannot be half-built. Every election on your profile carries both a document reference and the officer who signed it. A waiver holding one without the other is rejected the moment the object is built, so it cannot exist in memory, let alone reach a dial, and the database refuses the row from the other end. While your attestation is unsigned, your contacts screen undialable with the missing document named, so the dial waits for the signature instead of going out on an open question. What a signature gets you. A named owner for the screening behind every call, on the record, dated to the moment the call was placed. Whether the call was one you were entitled to place is yours to determine with your own advisors. The four controls enforced in code are unchanged by this and run on every call either way. The compliance engine is the page that owns them, and the evidence file is where an election lands. ## A date on this page is a date you can plan against An entry earns its place by shipping. The capability behind it is in the product on the day the entry appears, which is what makes this list worth reading before a renewal, a procurement review, or a conversation with your own team about what changed. Every capability sentence in an entry traces to the specification that owns it, and the reference is recorded with the entry so you can re-check a claim as the product moves underneath it. Where a change has a number, the entry links the page that publishes it rather than restating it: therate card owns the money, and thecompliance engine owns the controls. ## Read it, then go run it yourself A sandbox account is free, and everything above is the same build you get in it. --- # AI receptionist vs answering service | Vocapable > One monthly price for inbound answering, however busy the month gets. What a reserved connection buys, what it costs, and who decides what your caller hears. URL: https://www.vocapable.com/compare/ai-receptionist-vs-answering-service Compare · Inbound answering # One price, however busy the month gets Per-call human answering charges you for arrivals, so your best month costs the most. A reserved connection is one monthly price, held for you whether the phone rings once or all day. And every step a caller walks through is a step you wrote down yourself. ## Two things to buy, defined Two ways to buy someone answering your phone. The difference shows up on the invoice in your busiest month, and again in who decides what a caller hears. What Vocapable sells ### A reserved connection A monthly price for one concurrent inbound answer, held against your own traffic from the moment it is paid for. Buy more and more callers can be in an agent conversation at the same instant. The price does not move with how many people ring you, and the steps a caller walks through are the ones you wrote down. The other model ### Per-call human answering, by definition A price per call, per message, or per minute of a person’s time. What you pay tracks how many people ring you, so a busy month and a quiet month cost different amounts, and what the caller hears is whatever the brief you wrote said. ## Seven axes a buyer can actually check Every one of these is structural: who owns what, who writes what, and what the bill is a function of. They are the parts you can check for yourself before you spend anything. The right-hand column is the definition of per-call answering. It names no vendor and reports no vendor’s terms. Everything in the middle column is a first-party statement about how Vocapable works, checkable against the agreements published on this site. ## What reserved means here Exactly what the word covers, in the words the decision record uses, so the answer you get on a call is never smaller than the answer on this page. ### Held against your own traffic Read straight off your subscription, live from the first connection you pay for, with nothing to configure. Your own dialer at full pace cannot take the connection you bought, because the inbound and outbound counts never draw from one another. ### Yours, rather than fleet-wide The reservation is against your own account, which is the part you can plan around. The platform-wide inbound reserve is a separate thing and it ships at zero, so “reserved” here means reserved against your own traffic and nothing wider. ### One connection carries a much bigger flow A keypress menu, a ring to your own people, a forward and a spoken message claim no reserved capacity at all. Only the agent leg does. That is why one connection can carry an answering flow far larger than the concurrency it bought. ### The caller lands where you chose When every connection you bought is busy, the caller advances to the next step you wrote down, usually your own voicemail. Nobody sits listening to hold music, because what happens next is the thing you configured to happen next. The whole answering policy, the six step types, both chains, and how the open hours you set in your own timezone differ from the compliance calling windows computed against the person being called, is on the Receptionist page. ## What you pay, and when you pay it The plan $250/mo Includes 1 concurrent inbound connection, live from the moment it is paid for. Sold on its own, or attached to an outbound subscription for the same monthly money. Each additional connection $150/mo In any quantity. Each one lets one more caller be in an agent conversation at the same moment. And that is the invoice An answered call carries no per-call-attempt fee and no per-minute AI charge: both are charged on the calls an account places. So the plan and the connections you chose are the whole of what you owe us. A connection is prepaid a full month when you buy it, whatever day of your billing period that falls on, and it goes live the moment that charge succeeds. Buy one a few days before your subscription renews and it pays its prepaid month and then renews. Checkout tells you the exact amount and the exact renewal date before you agree to anything. The other direction is the one worth having in writing: a receptionist you already hold is not switched off over a single failed payment. A card that fails in month four leaves it answering while the dunning process runs, because the caller it would strand is your caller rather than you. The whole rate card, including the outbound plans a receptionist can be attached to, is on the pricing page. ## Three things that stay in your favour The parts that still matter a year in, written down before you buy. - The number stays yours. The receptionist answers on your own Twilio subaccount, on a number you already publish or port in, pointed at a signed webhook. Connecting that account is the longest step in getting started, you do it once, and it is the reason the number, the registrations and the caller-ID reputation stay with you. Your own carrier bills you for the carrier service directly, at the rates you negotiated. - Answering capacity your dialer cannot take. Receptionist is bought as inbound answering, and the two capacities never draw from one another, so a campaign at full pace leaves the connection you bought to answer with exactly where it was. Want the same account placing calls too? An outbound plan sits alongside it for the same monthly money. - What you can hold us to is the mechanism. A connection reserved against your own traffic, and an answering policy that advances the caller to whatever you configured next. The policy is yours to compose and review in the sandbox before you spend anything, and the whole agreement is published here so you read it before you sign it. The trust close ## An answered call leaves a record either way The four controls that have no off state run on a call you answer exactly as they run on a call you place, and nothing in your answering policy can reach them: the AI disclosure plays before the first model token, an opt-out is detected and written to your internal do-not-call list, that list suppresses, and numbers are normalized before anything reads them. The outbound-only checks are recorded as not applicable, each with the reason it does not apply, so an answered call reads in your evidence as an answered call rather than as a call that slipped its checks. Every entry names why it is there. You and your advisors decide what the law requires where you operate. The platform applies the values in force on every call and writes down what bound it, so the record is waiting for you when somebody asks. What an answered call records - Outbound screening, not applicable: the caller initiated the contact. - Outbound consent, not applicable: someone reaching you does not grant consent to call them. - Calling windows, not applicable: nothing was dialed. - Frequency caps, not applicable: nothing was dialed. - Outbound suppression, not applicable: it suppresses outbound contact, not a call somebody placed to you. The whole shape of a call’s record, and the division of responsibility around it, is on the Trust Center. ## Questions people ask first Because they are two different purchases, and the arithmetic differs before anything else does. A per-call price scales with how many people ring you, so a busy month and a quiet month cost different amounts. A reserved connection is one monthly price, held for you whether or not anyone rings, so arrivals do not move it. You are the only person who knows your call volume, and with a reserved connection your busiest month costs what your quietest one did. The caller advances to the next step you wrote down, which is usually your own voicemail. Nobody sits listening to hold music, because what happens next is whatever you configured to happen next, which is why the fallback step is worth writing carefully. Buying another connection is what lets one more caller be in an agent conversation at the same moment. It keeps answering. A connection you already hold is not switched off over a single failed payment; it stays live while the dunning process runs, because the caller it would strand is your caller rather than you. The other direction is the one worth having in writing too: a new connection is prepaid a full month at purchase and goes live the moment that charge succeeds, and checkout tells you the exact amount and the exact renewal date before you agree to anything. Yes, and it is the point rather than the paperwork. The receptionist answers on your own Twilio subaccount and the number you already publish, so the carrier relationship, the registrations and the caller-ID reputation stay yours. Connecting and verifying that account is the longest step in getting started, and you do it once. Nothing beyond the plan price and the connections you chose to buy. An answered call carries no per-call-attempt fee and no per-minute AI charge, because both are charged on the calls an account places. Your own carrier invoices you for the carrier service directly, so a Vocapable invoice carries no carrier line: no carrier minutes, no number rental, no markup. The busiest month you have all year costs what the quietest one did. ## Compose the answering policy before you buy the connection The sandbox gives you the whole builder: no card, no carrier account. Walk through exactly what a caller would hear before any number points anywhere. When it sounds right, Receptionist is one purchase in Billing. Related: the Receptionist product page, and telephony ownership for why the number stays yours. --- # BYO carrier vs resold minutes | Vocapable > Your carrier rate is the rate you pay. Every Vocapable production call leaves an account you own, on numbers you keep, and our invoice carries no carrier line. URL: https://www.vocapable.com/compare/byo-vs-resold-telephony Compare · Telephony ownership # The minutes stay on your bill, at your rate One model puts your calls on a carrier account you own: your numbers, the rates you negotiated, the caller-ID reputation compounding on an asset you keep. The other puts them on the platform’s account and hands you back a line item for the minutes. Vocapable ships the first one. This page is what that difference is worth, compared as models, with no company named on either side of it. ## The two models, defined Before anything can be compared, the thing being compared has to be a structure rather than a company. These are the two ways the carrier relationship can be arranged underneath calling software, and everything further down this page falls out of which one you are buying. Model one ### Bring your own carrier You hold the account with the carrier. You bought or ported the numbers, the registrations sit in your name, and the per-minute rates are the ones you negotiated. The software runs on that account, and your carrier bills you for the carrier service directly, at your rate. This is the only model Vocapable ships, and it is the reason our per-minute rate is as small as it is. Model two ### The resold model, by definition The platform holds the account with the carrier and gives you access to it. Numbers are assigned out of the platform’s inventory, the registrations sit in the platform’s name, and the minutes come back to you as a line on the platform’s invoice at whatever rate the platform sets. That is the definition of the model rather than a report on any company. ## The axes that actually differ Every row below is the same question asked about a different thing: who holds it. Read down the middle column. That is the column you keep on the day you change software. The third column is the definition of the resold model; it names no vendor and reports no vendor’s terms. The second column is a first-party statement about how Vocapable works, and every row of it is checkable against the agreements published on this site. ## What the difference is worth None of what follows is a preference. Each one falls out of the structure the way a sum falls out of its terms. ### Your Twilio rate is the rate you pay Twilio invoices you for the carrier service directly, at the rates you negotiated, so a Vocapable invoice carries no carrier line at all: no carrier minutes, no number rental, no markup, nothing to reconcile against a bill we never saw. What our invoice carries is your plan and your AI usage. ### Negotiate a better rate and you keep the saving The published per-minute rate prices the AI session: the model, the speech, and the session time on our own hardware, never the carrier leg. So the day your carrier gives you a better number, that saving lands in your pocket the same month, and nothing about our price changes. ### The reputation compounds on numbers you keep Answer rates follow the number, and the number is yours. Whatever your caller ID earns over a year of calling, it earns on an asset sitting in your own account rather than in someone else’s, and it does not reset because you changed software vendors. ### Our numbers stay out of your traffic The platform’s own carrier account is walled off from production calling. The only thing its numbers are ever permitted to carry is a paid test call to a destination proved through code verification; any other destination is refused at the dialability gate, and the refusal is written into that attempt’s evidence. Enforced in code, not promised on a policy page. Your own contract says the last one too, two clicks from here. MSA §6.2 makes numbers under Vocapable’s carrier account available solely for paid, low-volume test calls and states that the Service enforces the restriction technically; the Platform Test Calls Addendum adds that they are paid usage: metered, monitored, logged and volume-capped. ## One setup step, and the asset is yours for good You connect the account once, and everything it buys you outlasts the software contract. Here is the whole of that step, written out before you start rather than after you have signed. ### Connect the account once You connect and verify your own Twilio subaccount before the first production dial, and you point the number at the signed callback URL the platform hands you. It is the longest step in going live, you do it once, and everything after it is software. ### The registrations sit in your name They renew on your schedule and the rates underneath them are the ones you negotiated. Owning the account means its paperwork and its failures surface on your own account, in front of you, rather than inside somebody else’s support queue. ### Nothing to unwind later The account was yours the whole time, so changing software is a matter of pointing your numbers somewhere else. No porting request to file with us, no carrier contract of ours to exit, no reputation left behind. The mechanics of connecting an account, what verified means, and what the platform checks before it answers an inbound call, are laid out on the telephony page. The trust close ## You own the line, and the record of what went down it Every production dial leaves your account, aimed at a list you chose. You and your advisors set the standard that list has to meet, and from there the platform does the work. It applies the calling windows, the frequency caps, the consent standard and the recording-consent class in force on every attempt, and records the values that bound the call. The evidence of what was enforced is yours, on your side of the line, ready to produce. And the floor does not move because the account changed. The AI disclosure plays before the first model token, an opt-out writes your internal do-not-call list at once, that suppression is applied at the batch run and again at dial time, and every number is normalized. Same on your carrier account as on ours, same on every plan, with no flag that turns any of it off. The rest of the division of responsibility is on the Trust Center. The money half of this comparison, what a published rate card gives you that a figure on request cannot, is its own page: a published rate card vs a quote. The card itself is on the pricing page, and it is the whole of it. ## Questions people ask first Yours, on every plan and every tier, from the first dial. Production traffic rides the Twilio subaccount in your own name, on numbers you bought or ported, under registrations held by you. Numbers under Vocapable’s own account are permitted to carry a paid test call to a destination proved through code verification and nothing else, and that restriction is enforced at the dialability gate rather than asserted in a policy document. Nothing for the carriage. Your own carrier invoices you for the telephony directly, at the rates you negotiated, so a Vocapable invoice carries no carrier line at all: no carrier minutes, no number rental, no pass-throughs, no markup, nothing to reconcile against a bill we never saw. What our invoice carries is your plan and your AI usage, which is the whole of what you owe us. No, and that is the tell. The published per-minute rate prices the AI session: the model, the speech, and the session time on our own hardware, rather than the carrier leg. It does not move with the carrier account underneath it, so when you compare two per-minute figures, check whether the other one has carriage folded into it. They stay exactly where they already are. The subaccount is in your name, the numbers were bought or ported by you, and a year of caller-ID reputation is attached to numbers you keep, so leaving is a matter of pointing them somewhere else. There is no number to port back from Vocapable and no carrier contract of ours to exit, because the account was yours from the first dial. You connect and verify your Twilio subaccount, and you point your number at the signed callback URL the platform hands you. That is the whole of it. It is the longest step in going live, you do it once, and everything after it is software. The mechanics are written out on the telephony page so you can read them before you start rather than after you have signed. --- # A compliance floor vs a setting | Vocapable > A control nothing can switch off is a different thing from a setting you enable. Vocapable runs four of them on every dial, on every plan, at no charge. URL: https://www.vocapable.com/compare/compliance-floor-vs-compliance-flag Compare · Controls and settings # A control you cannot switch off is a different product A setting is something someone can turn off: an administrator, a plan tier, a support ticket, a spending limit. A floor is something nobody can, because there is no value anyone could write down that would express it. Vocapable runs four of them on every dial, on every plan, at no charge. Every call your account places arrives with a record of which ones ran and what bound it, which is what answers the question a year later without anybody having to remember it. The two models ## Two words that look like synonyms on a feature grid On a grid of checkmarks, “AI disclosure” and “opt-out handling” appear as rows in both columns, and a checkmark cannot say which kind of thing is behind it. The difference is not how well a control works. It is whether an off state exists. ### A floor, by definition A floor is a control with no off state. Not a control somebody promised not to turn off, and not a control that requires two approvals to turn off. A control for which no off state exists to be reached. The test is not who holds the switch. It is whether a switch was ever built. ### An opt-in setting, by definition An opt-in setting is something someone can turn off: an administrator, a plan tier, a support tool, a spending limit. That is not a criticism of any product, and it is not a claim about any product. It is what the word means, and it is the whole of what the right-hand column below describes. Enforced in code · every plan · every entry point ## Four controls, and nothing turns them off They run in every mode, for every tenant, on every plan, through every entry point on campaign dials, ad-hoc API calls, and agent test calls alike. They are not settings, and no API field, plan tier, feature flag or support tool reaches them. ### The AI-disclosure opener Compiled by the policy engine and played before the model produces its first token. The ordering is an interlock rather than a convention: reaching the model early raises instead of dialing, and if the opener cannot be played, the call ends. ### Opt-out detection, and the write that follows A deterministic phrase list and a keypress detector run on every final transcript. Either one firing hands control to scripted code, writes your internal do-not-call entry mid-call, and cascades the suppression across every campaign in your workspace. ### Internal-DNC suppression, twice Your own suppression list is applied when the list is prepared, and read live again at dial time rather than from cache. An opt-out taken during one call suppresses the next dial. If that read is unavailable, dialing halts rather than proceeding on a stale answer. ### Phone normalization Validation to E.164, short-code rejection, and dedupe, run as a stage of the pipeline rather than as a pre-step, so it leaves a result on the record like every other stage. ### Unrepresentable, not merely forbidden This is the whole argument of the page, and it is an engineering statement rather than a posture. A tenant’s compliance mode can only name checks drawn from a fixed waivable set, and these four are not members of it. “Waive the disclosure opener” is not a value anyone can construct, express in JSON, or store: not by an admin, not by a plan, not by a support tool with database access. Making a state unrepresentable is stronger than validating it away, because there is no code path left where a bug could permit it. ### Never metered, never paywalled There is no plan flag, allowance, overage, or spending-limit interaction that can reduce them. An account that has reached its spending limit stops dialing; it never keeps dialing with the floor off. What a paid tier buys here is advanced policy packs and audit exports, never safety itself, which is why therate card is a separate question from the floor. The axes ## The questions a checkmark cannot answer Each row is structural: what exists, who can reach it, and what is written down afterwards. Read the middle column as the answer you can give when your own leadership asks what stops a bad call from going out. The third column is the definition of a setting. It names no vendor and reports no vendor’s terms, because a claim about somebody else’s product would have to be re-verified against their live page on the day you read this, and nothing on this site is published that way. The record you keep ## Every call arrives with its record already written You keep the list arrangement you already trust and run it on your own contacts. Every production account signs one attestation naming that arrangement, the officer who signed it and the date, and the platform carries it from there. Each stage that attestation covers rides into every call’s evidence file as your election, with the signer named beside it. That is a record an auditor can read months later without asking anybody a question. Checked and clear, checked and suppressed, and elected away under signature stay three different rows, and nothing in the file lets them look alike. The same mechanism covers your paperwork gap. A check that is simply missing, with no election behind it, is a bug rather than a waiver: an account carrying no attestation for the stages it runs upstream screens its contacts undialable, with the missing document named, so a dial waits for a signature instead of going out on an unanswered question. And you can check all of it before you spend anything. The whole agreement set, including your Telecom Compliance Addendum, is published on this site rather than described, so what you read here and what you sign are the same document. The division ## We do the enforcing. You keep the evidence. Alongside the floor, the platform applies quiet-hours windows, frequency caps, consent-basis standards and a recording-consent class at the gate, and records the values it evaluated on every attempt. You and your advisors decide what the law requires where you dial. What you get from us is the enforcement, plus the evidence of it, in a file you can produce yourself. Those values are also the reason the floor names four things and stops. Quiet hours and frequency caps run on every dial too, and they are deliberately not floor controls: they are values, evaluated against a record, a clock and a counter, and a value is a different kind of thing from a control with no off state. Counting them in would make the list longer and less true. The mechanism behind all of it lives on the compliance engine page. What these controls do is block the calls that must not happen and write down what bound the ones that did. That is the part of the job you stop paying people to remember, on every call, every day, whether or not anyone is watching the queue. The trust close ## What decides it is what the record can tell apart Every dial attempt, including the ones the platform held back, produces an exportable record of which checks ran, which were elected away and on whose signature, and why the call was allowed or refused. A check that cannot complete blocks the call. Uncertainty never resolves in favour of dialing, and a stage that did not run is never recorded as a stage that found nothing. Which is the shortest way to state the axis this page is built on. A record that cannot tell checked, clear, and nobody looked apart is a log rather than evidence. And a control that can be switched off has to be proven off-or-on for every call you are ever asked about, which is a question a settings history answers and a type system does not have to. Tenant isolation is enforced by the database itself rather than by application code remembering to filter, and the service refuses to start if that isolation is misconfigured. The rest of the posture, control by control and duty by duty, is on the Trust Center. ## Questions people ask first No, and not because a policy forbids it. A tenant’s compliance mode can only name checks drawn from a waivable set, and these four are not members of it; the unwaivable stages are derived by set difference rather than listed, so they cannot be omitted by an edit. Waiving the disclosure opener is not a state anyone can construct, express in JSON, or store. That is stronger than a promise, because there is no code path left where a bug could permit it. A file that was written on the day of the call, by the system that placed it. Every attempt your account made, including the ones the platform held back, carries its own policy snapshot: which checks ran, which were elected away and on whose signature, the consent record relied on with its content hash, the calling window and the attempt cap evaluated in the contact’s own local time, and the pinned agent version. It leaves as CSV through the exports API, filtered to the contact and the dates you were asked about. Nobody on your team has to remember anything. Never, on any plan or class. The local floor stages are excluded from the metered lookups by construction rather than by a billing rule someone remembers to apply, so there is no configuration in which the floor is charged for or switched off to save money. An account that has reached its spending limit stops dialing; it never keeps dialing with the floor off. A waived stage renders as an election: elected away under a stored attestation, with the officer who signed it named. A stage that found nothing renders as a check that ran and came back clear. Those are different rows and they are never allowed to look alike. A check that is simply missing, with no election behind it, is neither one: it is a bug rather than a waiver, and an account carrying no attestation for the stages it runs upstream screens its contacts undialable with the missing document named. No. The platform applies the policy values in force (quiet-hours windows, frequency caps, consent-basis standards and the recording-consent class) and records the values it evaluated on every attempt. That is an enforcement mechanism, not a determination of what the law requires where you dial. Vocapable does not guarantee that any call is lawful. ## Read a policy snapshot before you decide Open a sandbox account and read the evidence from a run you set up yourself. The floor runs there exactly as it runs in production, and the elections render exactly as they would on a live account. --- # A published rate card vs a quote | Vocapable > Every figure Vocapable bills is printed on one page, and your carrier minutes stay on your own bill. Do the whole sum today, before you talk to anyone here. URL: https://www.vocapable.com/compare/published-rate-card-vs-per-minute-quote Compare · How the number reaches you # A published card, or a number you have to ask for A per-minute headline is one input. Your bill is that input times your minutes, plus whatever the headline left out. This page prints every figure on our side of that arithmetic and says exactly what our per-minute number prices, so you can do the sum this afternoon instead of discovering it a quarter later. ## Two ways a price reaches a buyer These are models rather than companies. One of them is what we do; the other is defined here as a shape, so the comparison is about structure and you can apply it to whatever else is on your desk. What we do ### A published rate card Every figure sits on a page you can read without talking to anyone: the monthly price of each plan, the concurrent sessions it includes, the price of another one, the per-call-attempt fee, both AI-minute rates, the volume ladder and the floor it stops at. You can compute a month before your first call, and the figure you compute against is the figure we invoice. The other model ### A per-minute quote, by definition One headline figure, given on request. Whether it covers the carrier leg, the platform fee, the concurrency, the model cost or the transcription is a property of that particular quote, and the buyer usually finds out by adding up invoices. That is what a quote is: a number without a published container around it. ## Six axes a buyer actually has to decide on Every one of them is structural: where a figure lives, what it contains, what moves it, whose invoice it lands on. Each row is one you can decide for yourself, today, with nothing but this page and last month’s phone bill. The right-hand column is the definition of the model. It names no vendor and reports no vendor's terms: a comparison built on somebody else's live pricing page goes stale without anyone noticing, and we would rather publish our own numbers than characterize theirs. ## Every figure we bill, on one screen Not a summary of the card. The card. Each number below is read from the same module the pricing page reads, so this page cannot quietly disagree with the one you buy from. Outbound plans, per month A class is a ceiling on the compliance weight your calling carries rather than a channel, and the higher ones include the work of the ones below. Each additional concurrent session is $200 per month, in any quantity. Usage, on the calls your account places - $0.005 per call attempt, on every outbound plan alike. - $0.13 per minute of AI usage, month to month. - $0.09 per minute of AI usage on a signed 3-year commitment: the same monthly base, non-cancellable, all sales final under a signed non-refund waiver. - $250 per month for inbound answering, including 1 concurrent inbound connection, with each additional one $150 per month. It answers rather than places calls, so neither charge above applies to it. ### The ladder, printed rather than negotiated For every $10,000 of total monthly spend (plan base, session add-ons, call fees and AI minutes together, rather than minutes alone) the AI-minute rate steps down $0.01, to a floor of $0.07 per minute. A step applies mid-month, going forward from the moment your spend crosses the threshold, and the meter resets at the start of each billing month. It stacks with the committed rate, so both tables end at the same floor. #### Month to month #### 3-year commitment ## What our per-minute number prices It is the AI-usage rate: the conversation, not the line it travels on. The model, the speech, and the session time on our own hardware. Every production call leaves a Twilio subaccount you own, and Twilio invoices you for its carrier service directly at the rates you negotiated, so a Vocapable invoice carries no carrier line at all: no minutes billed as carriage, no number rental, no markup. That is why the rate holds steady when the carrier arrangement changes. It prices the AI session rather than the minute a carrier sold, so it is the same figure either way. It also means a per-minute figure that carries the line and one that does not are different quantities, and comparing them as though they were the same gets the sum wrong in our favour. We would rather you got it right. The carrier half of this has a page of its own: telephony ownership, and the model comparison beside it, whose account the call leaves on. Where each charge comes from - Yours The carrier minutes, the number rental and the registration pass-throughs, invoiced to you by your own carrier, on your own agreement with them. - Ours The plan, and the AI usage rated against it. That is the list, and every figure in it is above on this page. Two bills from two companies, each for the thing that company actually did. The arithmetic a buyer has to do is smaller when nobody is holding a minute of yours in their books on the way past. What a published card commits us to ## The number is the number There is no volume conversation that starts below what is printed, because the ladder is printed too, and its floor is printed with it. Whatever you compute from this page is what arrives on the invoice, and you never have to wonder whether the buyer before you got a better sentence out of a sales call. The estimate is built to lean the same way. The calculator on the pricing page takes your assumptions about your own list rather than supplying flattering ones of its own, and it deliberately leaves the volume discount out, because that discount brackets on the whole account’s realized monthly spend and a single campaign shape cannot know it. An estimate that came in under the invoice would be the one kind of wrong we are not willing to be, so it can only miss high. The card itself can move. What holds is that you read it here first, in full, and that a signed order keeps its rates for its whole term. Design-partner pricing, published so you can plan, and locked for the term of any order you sign. The trust close ## Nothing here is a figure we would not put in an Order Everything on this page is on the pricing page too, in more detail, from the same module, including the calculator that prices your own campaign shape against these rates. This page exists to say what the numbers mean, not to hold a second copy of them. Two things sit outside the arithmetic on purpose. The compliance controls are included on every plan and are never a paid tier, so there is no safety line in your total to compare against anybody else’s. And you and your advisors decide what the law requires where you dial; the platform applies the policy values in force and records what bound each attempt, so the record is yours to produce. The division of responsibility is laid out on the Trust Center. ## Questions people ask first On the pricing page, and it is the whole of it: the monthly price of each outbound plan, the concurrent sessions each one includes, the price of another session, the per-call-attempt fee, the AI-minute rate on either commitment term, the volume ladder and the floor it stops at, and the inbound answering plan with its per-connection price. Every figure on this comparison page is rendered from the same module the pricing page reads, so the two cannot disagree with each other. Your carrier minutes stay on your carrier’s bill, at the rates you negotiated. Our rate prices AI usage: the model, the speech, and the session time on our own hardware. Every production call leaves a Twilio subaccount you own, and Twilio invoices you for its carrier service directly, so a Vocapable invoice carries no carrier line: no minutes billed as carriage, no number rental, no markup. That is why the number on our card is as small as it is, and why it holds whatever your carrier arrangement does. The monthly plan base, the per-call-attempt fee on the calls your account places, and any concurrent session beyond the ones your plan includes. All three are published beside the minute rate rather than found later, which is what lets you compute a month before you buy one. The compliance controls are included on every plan and are never a paid tier, so there is no safety upsell waiting in that column either. It is built so that it cannot. The estimator takes your assumptions about your own list rather than supplying flattering ones of its own, and every one of them stays editable and visible. It also leaves the volume discount out on purpose, because that discount brackets on the whole account’s realized monthly spend and a single campaign shape cannot know it. Both choices bias the number against us, which is the only direction an estimate is allowed to miss in. This is design-partner pricing: published so you can plan against it, and locked for the term of any order you sign. The card can move; your signed rates cannot. What holds either way is the shape of the arrangement. The card is published rather than quoted, and no volume conversation starts below a rate the ladder already prints. Nothing per call, however busy the month gets. Receptionist answers calls rather than placing them, so there is no call attempt to charge a per-attempt fee against, and an answered inbound call carries no per-minute AI-usage charge either. A Receptionist invoice is its monthly plan price plus its additional connections, and that is the whole of it. --- # Compliance enforced in code | Vocapable > The compliance engine: four controls no plan or flag disables, one gate every dial passes, and an exportable evidence file for every call and every refusal. URL: https://www.vocapable.com/compliance Product · The compliance engine # Every call, on the record Every dial passes one gate, and every dial leaves a file: what ran, what your signature covered, what the agent said, and why the call was placed or refused, all of it reconstructible years later from the record alone. Four controls are enforced in code on every plan, for every tenant, through every entry point. No API field, plan tier, feature flag or support tool reaches them, because they are not settings: a compliance mode can only name checks drawn from a waivable set, and these four are not members of it. Waiving the disclosure opener is not a state anyone can construct, express in JSON, or store. Why this exists ## A demand letter asks one question What happened on this call, on this date, to this number. Most calling stacks answer it by hand, months later, out of logs that were never designed to be read by anyone outside the company. This engine was designed backwards from that letter. Three tenets govern every mechanism below, and they are worth reading before the mechanisms, because each one is the reason a mechanism has the shape it has. ### Fail closed When a compliance check cannot complete, the call does not happen. Only an affirmative pass authorizes a dial, and a stage that cannot answer yields an unknown that resolves to a suppression. ### Deterministic enforcement, probabilistic assistance A model may detect opt-out intent or sensitive content. Only deterministic code may enforce: block a dial, write a suppression row, play a disclosure. The line between those two verbs is the whole architecture. ### Everything leaves evidence Every decision must be reconstructible years later from the audit store alone, without access to the live system. A control nobody can prove ran is a control you cannot use in an answer. Enforced in code · every plan · every entry point ## Four controls, and nothing turns them off In every mode, for every tenant, on every plan, through every entry point: campaign dials, ad-hoc API calls, and agent test calls alike. ### The AI-disclosure opener Compiled by the policy engine rather than generated by the model: who is calling, on whose behalf, that the voice is an AI, and the recording announcement where it is required. You supply the values that fill it: company name, agent name, and an optional extra line that can lengthen the opener and never shorten it or replace a required element. The AI sentence is unconditional and cannot be prompt-injected away. It is played before the model produces its first token, and the ordering is an interlock rather than a convention: the model adapters are held behind a gate the call driver holds no reference to, so reaching the model early raises instead of dialing. If the opener cannot be played, the call ends. Writes: The opener’s exact text, and playback confirmation by media-stream offset. ### Opt-out detection, and the write that follows it A deterministic phrase list and a keypress detector run on every final transcript, with a model classifier for the phrasings the list misses. Either one firing hands control to scripted code, and the model does not get to argue. The phrase list and the closing line it speaks are module constants, not configuration. The suppression is written and confirmed before the call closes, a fixed graceful close is spoken, the attempt is disposed opt_out, and the suppression cascades across every campaign in your workspace with pending dials cancelled. Writes: The internal do-not-call entry, the optout.recorded event, and the write itself recorded as written, failed, or not applicable. ### Internal-DNC suppression, twice Your own suppression list is applied when the list is prepared, and read live again at dial time, never from cache. An opt-out taken during one call suppresses the next dial. If the store behind that read is unavailable, dialing halts rather than proceeding on a stale answer. The dial-time re-check is a separate read from the batch one on purpose. A list prepared this morning is a photograph; the gate needs the present tense. Writes: The stage result on the run, and the refusal reason internal_dnc on the attempt. ### Phone normalization Validation to E.164, short-code rejection, and dedupe, run as a stage of the pipeline rather than as a pre-step, so it leaves a result like every other stage. It is the one stage whose own suppression stops the ladder, and for a structural reason rather than a policy one: there is then no number left for the suppression list to be asked about. Writes: The normalized number and the stage result. ### Why these four, and not others They are precisely the acts a regulator attributes to whoever originated the artificial voice, and that is Vocapable, even when the call leaves your own carrier account, under your own registrations, billed to your own carrier. We compile the disclosure, we synthesize the speech, we decide whether the next dial happens. That is not a duty a contract can hand back, so it is not a setting either. It runs on every plan and is never a paid add-on; therate card is a separate question from the floor. The difference between a control with no off switch and a setting that has one isworth seeing side by side. ### Why quiet hours and caps are not on that list Quiet-hours windows and frequency caps are enforced on every dial too, and they are deliberately not floor controls. They are values, evaluated against a record, against a clock, against a counter, and a value is a different kind of thing from a control with no off state. Counting them in would make this list longer and less true, which is the easiest mistake on this page to make. The gate ## Every dial goes through the same function Campaign dials, ad-hoc API calls and agent test calls all evaluate through one dialability evaluator. There is no second implementation and there is no side door: the verified-own-number exemption a test call relies on is an input to that function, not a way around it. The evaluator is pure: it takes an already-gathered picture of the world and returns a verdict plus the evidence for it, which is what makes every branch testable, including the ones that only occur at 8:59pm in an ambiguous area code. ### Two refusal vocabularies, kept apart One says this number may not be called at all. The other says the number is fine but this moment, this cadence or this consent basis is not. Collapsing them would leave an evidence export unable to answer the first question a reviewer asks: was this number callable at all, or just not right now? internal_dnc · consent_insufficient · consent_scope_mismatch · blocked_state · quiet_hours · sunday_ban · no_lawful_window · frequency_cap · platform_test_only ### Only a pass authorizes a dial Every stage yields a pass, a suppression, or an unknown, and an unknown resolves to a suppression. A check that cannot complete blocks the call. Uncertainty never resolves in favor of dialing, and a stage that did not run is never recorded as a stage that found nothing. ### Authorization is linearized, not hoped for The worker locks the contact, then a transaction-scoped suppression subject for that exact number, samples its clock only once both locks are held, and reads the gate facts with that fresh time. A revocation that wins the race commits first; the dial worker then observes it, disposes the attempt, and never asks the carrier to originate. Carrier traffic begins only after the seize has committed. Consent ## Four classes, ranked, and checked at the gate Consent is a typed entity on the record, not a checkbox column somebody set to true during an import. It has a class, a scope, a source, a proof, and an expiry. ### What a consent record carries Append-only, and it holds the contact, the class, its scope (purposes, campaign categories, channels), its source, its proof with a content hash, the disclosure text shown where the class demands one, when it was captured, when it expires, and its status. Written consent expires on a schedule you may tighten and never loosen, and expired consent downgrades dialability rather than silently passing. ### Scope is enforced, not decorative Consent to appointment reminders does not authorize a sales campaign: a campaign whose purpose falls outside a contact’s scope is suppressed with consent_scope_mismatch. The classes are ranked against the standard in force for the call’s purpose, and a purpose nobody listed takes the strictest standard on file rather than the most convenient one. One more reason the consent record earns this much machinery: it is the part of the file that speaks directly to whether this person agreed to hear from you. Nothing else in the record does that job, which is why it is typed, scoped, hashed, and read again at dial time instead of trusted from an import. Revocation ## One revocation, one write path A revocation is one fact about a person, not a per-channel preference. Somebody who says stop calling me has not asked to keep hearing from you somewhere else. ### One transaction, four writes A mid-call opt-out, a keypress, an inbound stop request and your own API call all converge on the same transaction: the internal suppression entry, the optout.recorded event, the withdrawal of the consent standing behind the number, and the dial-time fast path. Not four writes that have to agree afterwards. One write, whose statement order carries the guarantee. ### Channel-free by construction The suppression table has no channel column, so both readers ask it the same question and there is no second lookup to fall out of step. A revocation whose covered set cannot be resolved covers everything: it over-suppresses rather than under-suppresses, which is the only direction this is permitted to err in. ### And it cannot be given an expiry The endpoint refuses one rather than honoring it. Two of those writes are irreversible: the consent ledger is append-only and the dial-time flag is never re-derived, so a time-limited revocation would destroy the consent evidence permanently and then hand the number back as dialable. ### On the regulation, as two separate facts 47 C.F.R. § 64.1200(a)(10) requires a revocation received through any medium to be honored across every channel the caller uses. The cross-channel portion becomes mandatory on 2027-01-31, after two delays. Vocapable propagates today rather than on that date, and the date itself is recorded as evidence and deliberately not branched on, because gating the behavior on it would mean shipping the non-propagating version first. That rule allows ten business days to honor a revocation. Ours lands in the same transaction that records it. Those are two different numbers, and neither one is ever quoted as the other: the statutory bar is what a record gets measured against, never a schedule anything is allowed to drift toward. The attestation ## One signature, and every record can answer for itself Keep the screening arrangement you already run and the pricing you negotiated for it. Your posture is never locked to a data vendor we picked, and you never pay twice for a lookup you already buy. Before your first production dial you sign one attestation naming that arrangement. It does not sit in a folder afterwards. Its reference and its signer are written onto your compliance profile and ride into every attempt from then on, so each stage your own arrangement covers arrives in the file as a named election with the document and the person behind it. Six months later, when somebody asks who screened this number, the answer is on the call rather than in an email thread. The platform will not let that election be half-written. A waiver carrying a reference and no signer is rejected the moment the object is built, so it cannot exist in memory, let alone reach a dial. The database holds the other end of it: a profile carrying waived stages with no attestation reference is a row it will not accept. And an account with nothing signed does not screen clear either. Its contacts screen undialable, with the missing document named, because a missing check with no election is a bug rather than a waiver, and the gate treats it as one. So your file hands a reviewer three things they can act on: a name, a document, and a date. Follow any of them and you land on a person who signed. The record ## One call, one file Every attempt, connected or not, placed or refused, produces exactly one policy snapshot. A call with no attempt behind it carries its snapshot on the call instead. One home per call, never both, never neither. policy_snapshot atmp_01J9…K4T Product UI · synthetic data Synthetic tenant, real shape. The amber row is the point of the panel: a stage your own arrangement covers arrives as a signature, never as a silence. ### Two rules govern how this file renders Both are load-bearing, and breaking either one is treated as a defect of the highest severity rather than a cosmetic issue. A check that did not run must never read as a check that found nothing. And a value we applied must never read as a value we vouched for. So the file names the window, the caps, the consent standard and the recording-consent class that bound the call, and dates the rule pack each one came from. What the file settles is which values bound this dial. What they mean for your program is yours to determine with your own advisors. ### What else it carries The stage results and, for each stage you run upstream, the election with its attestation reference and signer. The consent record relied upon, with its content hash, or the exemption class instead. The quiet-hour and attempt-cap evaluations with the computed contact-local time. If the call connected: disclosure playback confirmed by media-stream offset, recording-consent status, the opt-out and transfer events, the disposition, and the suppression write recorded as written, failed, or not applicable, and that last one is an election rather than a softer failure, because a leg with no called party has no number to suppress. ### Kept, and kept honest Compliance evidence is retained at least five years, the four-year statute of limitations plus margin. It is not tenant-configurable and it is excluded from tenant-initiated deletion, because it is the platform’s defense record as much as yours. No database role holds UPDATE or DELETE on the audit log, so the application tier could not rewrite history even by bug; a correction is a new row referencing the one it corrects. ### Getting it out A phone number and a date range is the shape a demand letter arrives in, and it is also the shape of an export filter. Submit a job for the attempt rows and their snapshots, and another for a run’s screening results. A completed job provides a signed CSV download that expires with the artifact it points at. Every row carries the values that bound the call, the consent record it stood on, and the attestation covering each stage your own arrangement runs, with its signer and its date. The export itself lands in the audit log. Record production is a filter and a job, not a week of somebody’s life. Before the first dial ## Nothing launches until somebody signs off on the list Before a campaign can launch you see the run report: the funnel from uploaded records to dialable ones, per-stage suppression counts, the reason on every suppressed row, and the policy versions that bound the evaluation. The records the platform will not call come back as a downloadable export, with the reason beside each one. ### 1 · The run The floor stages plus whatever you elected upstream. A suppression short-circuits the stages that would have cost money to perform; the floor stages still run to the end, so every run’s evidence carries an internal do-not-call result. Suppression is not what stops the ladder. Cost is. ### 2 · The acknowledgment Launch is refused until an authorized user acknowledges the report, and that acknowledgment records who, when, and the totals exactly as they were seen. Nobody gets to say later that the numbers were different when they clicked. ### 3 · Staleness Results expire. A campaign paused past the staleness window re-runs incrementally before it resumes, rather than dialing on an answer that was true last month. Every stage you elected to run upstream appears in this report as waived, with its attestation reference and its signer, and never as a stage that found nothing. That is the same rendering rule the evidence file follows, applied at the point where you can still do something about it. The division ## What runs on every dial, and what stays in your hands Quiet-hours windows, frequency caps, consent-basis standards and the recording-consent class are applied at the gate on every attempt, and the values the gate evaluated go into the file. What the law requires of your program where you dial is yours to determine with your own advisors. Applying the values in force on every attempt, and proving afterwards which ones bound the call, is ours. Every plausible jurisdiction for a number is resolved and applied stricter-only, rather than one being picked: candidates come from the area code, the carrier locale and the address you supplied, and all of them bind, because a 305 number can live in Seattle. The layering can only narrow what is permitted. It never invents a window. Every production dial goes out on the Twilio account you own: your numbers, your registrations, your negotiated rates, and the caller-ID reputation attached to them. Lists you purchased, rented, or had appended are yours to call, under the Cold Outreach & Data Provenance Waiver you execute, which records that the data is yours under your own supplier agreement and that the consent basis, the jurisdictions and the responsibility for that program stay on your side. Scraped and unauthorized-extraction contacts are refused however they arrive. The four controls above are unchanged by that waiver, and an applicant who declines the AI-disclosure opener is refused, terminally. Every attempt either clears the gate or lands on the record as a refusal with the reason that produced it. A signature, on the attestation or on the waiver, records who carries which half of the work, so nothing in the division rests on an assumption. The security posture, the scope of service, and the whole published legal set live on the Trust Center. The trust close ## The claim that still holds on the day someone asks Here is what we will put in writing. Every call, placed or refused, is reconstructible from the record years later. The calls the platform can detect as ones you should not place do not go out. Both of those still hold on the day somebody asks you to prove them, which is the only day a compliance claim is worth anything. It is also why the difference between checked, clear, and nobody looked is a design rule here rather than a detail. A record that cannot tell those three apart is not evidence. It is a log. ## Questions people ask first No, and not because a policy forbids it. A tenant’s compliance mode can only name checks drawn from a waivable set, and these four are not members of it; the unwaivable stages are derived by set difference rather than listed, so they cannot be omitted by an edit. Waiving the disclosure opener is not a state anyone can construct, express in JSON, or store. On every call. You sign one attestation naming the arrangement you already run, and its reference and its signer go onto your compliance profile and into every attempt from then on. Each stage that attestation covers arrives in the file as a named election carrying the document and the officer who signed it, so the record answers who screened this number and on what date. A waiver missing either half is rejected the moment it is built, a profile carrying waived stages with no attestation reference is a row the database will not accept, and an account with nothing signed screens undialable rather than clear. The call does not happen. Every stage yields a pass, a suppression, or an unknown, and an unknown resolves to a suppression, and only an affirmative pass authorizes a dial. A check that cannot complete blocks the call, and uncertainty never resolves in favor of dialing. An account carrying no attestation for the stages it runs upstream does not screen clear: its contacts screen undialable, with the missing document named. In the same transaction that records it. A mid-call opt-out, a keypress, an inbound stop request and your own API call converge on one write path, which lands the internal suppression entry, the event, the withdrawal of the consent standing behind that number and the dial-time fast path together, or lands none of them. An opt-out taken during one call suppresses the next dial. Every channel it covers. There is one suppression row and it carries no channel column, so both readers ask it the same question and there is no second lookup to fall out of step. A revocation whose covered set cannot be resolved covers everything: it over-suppresses rather than under-suppresses, which is the only direction it is allowed to err in. It also cannot be given an expiry: the endpoint refuses one, because two of those writes are irreversible and a time-limited revocation would destroy the consent evidence permanently and then hand the number back as dialable. 47 C.F.R. § 64.1200(a)(10) requires a revocation received through any medium to be honored across every channel the caller uses, and the cross-channel portion becomes mandatory on 2027-01-31, after two delays. Vocapable propagates today rather than on that date. The date is recorded as evidence and is deliberately not branched on, because gating the behavior on a date would mean shipping the non-propagating version first. The platform applies them for you. Quiet-hours windows, frequency caps, consent-basis standards and the recording-consent class are enforced at the gate on every attempt, in the contact’s local time, and the values evaluated go into the file. Calling-policy configuration is not available in the portal or API, so nobody on either side loosens a window by accident. What those values mean for your program where you dial is yours to determine with your own advisors; applying them on every attempt and proving it afterwards is ours. No, on any plan or class. The two local floor stages are excluded from the metered lookups by construction rather than by a billing rule someone remembers to apply, so there is no configuration in which the floor is charged for, or switched off to save money. You do, with your own advisors. Our half is enforcement and evidence: the gate runs on every attempt, the calls the platform can detect as ones that should not go out are stopped before the carrier is asked to originate anything, and every attempt leaves a file. A signature, on an attestation or on the cold-outreach waiver, records who carries which half, so nothing in the division rests on an assumption. ## Every call, on the record. Included, on every plan. Open a sandbox account and read a policy snapshot from a call you ran yourself. The floor runs there exactly as it runs in production. --- # Who to email, and what you get back | Vocapable > One written address for complex rollouts, Enterprise conversations, account support, security questionnaires, and published agreements. URL: https://www.vocapable.com/contact Contact # One address, four routes Everything reachessupport@capstralabs.com. Write once and you are already in the right place. The four routes below differ by what happens next, not by where the mail lands. ## Sales and production The free sandbox and standard production path are self-service. Contact us for a complex rollout, a technical conversation, a security questionnaire, or custom Enterprise terms. ## Support Anything about an account that already exists. Write from the address on the account where you can, because it is the fastest way to be sure we are looking at the right tenant. ## Security Security questionnaires are answered directly and in writing, about the controls themselves. The control descriptions are published on the Trust Center rather than held back until somebody asks, so most questionnaires can be answered before you send one. ## Legal The agreement set is published in full, including the documents the master agreement incorporates by reference. Redlines and questions about any of it come to the same address. Vocapable works US business calling, in English. Every route above ends in writing, so you keep the record of what was said and so do we. ## Or just try it The sandbox costs nothing and answers most of the questions people write in about. --- # A CRM built for one job: the phone | Vocapable > A call workspace for contacts, outcomes, caller requests, and the evidence your team needs to act on each conversation. URL: https://www.vocapable.com/crm Product · The CRM for phone outreach # Every call, on one record Open one contact the morning after a campaign and you have the whole answer: who you reached, what they said in their own words, which campaign dialed them, and who asked you to stop. Every field of it is a/v1 call, so it belongs to you whether you open the dashboard or never do. A person checks their phone and smiles as campaign outcomes land (Illustrative footage.) ## From a dial to a record somebody can open - 1 The dial is authorized on the record Before anything rings, the gate reads this contact’s own row: internal do-not-call state, consent class, jurisdictions, and the callee’s local clock. The verdict and everything it read are kept on the attempt. - 2 The call becomes usable evidence Authenticated /v1 call reads return ordered turns and, where available, the transcript or recording for that call. Nobody retypes a call into a CRM at the end of the day. - 3 The attempt is disposed One value from a pinned vocabulary of fifteen: completed, answered but incomplete, voicemail, no answer, opted out, refused by policy, and the rest. Fifteen values means your reports add up. - 4 It lands on the contact Beside every other attempt anyone in your workspace ever made to this person, whichever campaign made it, whichever agent ran it. - 5 And it leaves again For enabled events, signed call.ended delivers the disposition, timing, record identifiers, and evidence availability at call completion. Structured surveys and shift confirmations have their own configured signed outcome events; their resources and supported CSV record types remain available through /v1 and export jobs. Dispositions are append-only. When your reviewer disagrees with the agent’s label, the review lands as its own row and the record says which verdict came from whom. ## Built for the fields you actually dial on A general-purpose CRM has a phone field. This one has a timezone, because the calling window is computed in the callee’s local clock; jurisdictions, because the policy resolves against them; and a consent class, because the gate reads it before every attempt. Your own columns come along too, kept as custom fields the API reads and writes. Tags and lead status are yours to move around freely. Stopping a call is a separate field with its own write path, and the screen where you set a tag says so, so nobody on your team has to guess which one holds a stop request. What one contact carries - Phone in E.164, name, company, title, email - Timezone and jurisdictions, the two facts the calling window is computed from - The best consent class held, and every consent record behind it, with its source and capture date - Internal do-not-call state, read from the list itself rather than a cached flag on the row - Tags, lead status, owner, and every list this contact belongs to - Custom fields, carried in from your CSV columns and your API writes - Notes in your own words, each one optionally pinned to the call it came from - Follow-up context your team records on the contact One person, open ## The whole history, and the state that governs the next dial Campaigns are how work gets scheduled. People are how it gets remembered. Open a contact and the attempts arrive in one column no matter which campaign placed them, with the consent and suppression facts sitting right beside them. contact ctc_01J9…H3D Product UI · synthetic data Dale Whitcomb · Whitcomb Freight Two campaigns, one column. Notes sit on the record beside the calls, and the call log filtered to this contact carries the full history behind them. dialability ctc_01J9…H3D Product UI · synthetic data Evaluated over this contact’s own record: the internal do-not-call flag, the consent class, the jurisdictions, and the callee’s local clock, against the policy in force at that moment. The second panel names the inputs the gate read and the verdict it returned, which makes it a record of the check rather than a legal clearance. In and out ## A system of record with doors on both sides Import contacts, follow calls and caller requests in one workspace, then move supported records through CSV exports, receive signed call and configured-outcome webhooks, or fetch a /v1 read. ### CSV in, with a receipt Map your columns to the fields, sign the consent attestation the import requires, and get back counts you can act on: accepted, rejected, skipped. Every rejected row comes back with its row number, its reason code, the column and the value that failed, as JSON or as a CSV you hand to whoever owns the spreadsheet. ### Bulk actions, one transaction Up to a thousand contacts at once: add to a list, remove from one, set lead status, suppress, delete. Suppress is not a flag write. It takes the same revocation path a single opt-out takes, so the do-not-call entry, the event, and the withdrawal of the consent behind that number all land together or none of them do. ### CSV exports, selected for the job Choose a supported record type: contacts, attempts, calls, survey responses, shift confirmations, appointments, do-not-call entries, scrub results, usage events, or caller requests. Add filters and columns, then a completed job provides a signed CSV download for its time-bounded window. ### Webhooks with a delivery trail The call-completion event and the outcome events you enable—including survey.completed and shift.confirmed—reach the endpoint you register, signed with a secret you control. Inspect the delivery history and redrive an exhausted delivery when your endpoint recovers. ### Caller requests, ready to work Messages, tickets, and callbacks an agent takes during a call enter a native Inbox, where your team can acknowledge, assign, or resolve the next step. Keep your existing system of record If your stack is already the system of record, drive/v1 directly with your own Twilio, create each contact as you dial it, and receive configured signed call-completion and structured-outcome events on your webhook. The workspace and API use the same engine, rates, and policy controls, so your team can keep the system it already runs while adding voice operations. The headless version The trust close ## Every call leaves proof, on every plan You choose who to call, on your own carrier account, and that choice stays yours. What the platform contributes is enforcement and memory: the AI disclosure played before the first model token, opt-out detection writing your internal do-not-call list, that list applied at the batch run and re-checked again at dial time, and phone normalization. Four controls no plan, flag, or support tool disables. On top of them, every attempt runs inside the calling window in the contact’s local time, under the frequency caps, against the consent standard and the recording-consent class, and the values that bound it are written onto the attempt. The basis you gave for calling that number travels with the contact into every attempt, so a record pulled a year later still answers why the phone rang. That is why a campaign run here is still answerable six months later: the panels on this page record what happened and refuse what you told them to refuse, and the evidence file behind each one holds up when somebody asks. ## What getting started actually looks like Sandbox first, then a contract, then production. The rate card and the $0 compliance floor are laid out on thepricing page. ### Start in the sandbox A free sandbox account gets you the whole workspace, no credit card. Import a list, work a contact record end to end, and hear the simulated demo calls. ### What you bring Your contacts and the consent basis behind them, by CSV or one at a time through the API, whichever your stack prefers. Purchased, rented and appended lists come in the same doors under the Cold Outreach & Data Provenance Waiver you sign, and every import carries a signed record of where the contacts came from. ### Production on your carrier When your workspace is ready, choose a plan, add a payment method, connect your own Twilio subaccount, and activate production from the portal. Your numbers, registrations, and carrier bill stay with the account you own. ## Somewhere for the answers to live. Included, on every plan. --- # Talk to sales | Vocapable > Talk to Vocapable about complex rollouts, technical architecture, and Enterprise terms. The sandbox and standard production path are self-service. URL: https://www.vocapable.com/demo Talk to sales # Build on your own. Bring us in when it helps. Thefree sandbox and the standard production path are self-service. Talk with us when you are planning a complex rollout, comparing Enterprise terms, or want a technical conversation about how Vocapable fits your operation. ## Tell us what you want to run For complex rollouts, Enterprise terms, and technical conversations. Send it by email instead. The link below reaches the same people. Prefer your own mail client? Write to support@capstralabs.com. It reaches the same people. Write to us and we will coordinate a time. A qualification call, start to handoff Simulated demo · AI voices - agent ▸ “Hi Jordan, quick heads-up: I’m an AI assistant calling for Northline Software. You grabbed our pricing guide yesterday. Got forty seconds?” - lead ▸ “Uh, sure.” - agent ▸ “What’s pushing you to look at tools like ours right now?” - lead ▸ “We keep missing inbound leads on nights and weekends.” - agent ▸ “That’s exactly the job. Give me a day and a window that work for you and I’ll put them in front of a Northline specialist with everything you just told me.” - lead ▸ “Thursday morning. After ten.” - agent ▸ “Thursday after ten, in your words. That’s on their screen the second we hang up, and they’ll come back to you to lock the time. Thanks, Jordan!” A scripted, simulated example in AI voices, opening with the same disclosure the product plays. Transcript above. Another scripted example plays on the home page, transcript beside it, nothing to fill in. ## What to expect ### A practical working session Bring your actual use case, lead source, list, and rubric. We can walk the product with you, pressure-test a rollout, or help you decide whether the self-service path fits. ### The compliance conversation up front Consent basis, list provenance, and which checks run where, answered before you sign anything. Every attempt runs inside a calling window computed in the contact’s local time, under frequency caps, against a consent standard, and the values that bound the call are written onto it. You will know which checks run where before you sign, not after. ### A straight answer on fit The agent works your leads and your people, captured or acquired, and every production dial leaves on your own Twilio account at the rates you negotiated. Cold outreach on a list you bought, rented or had appended is supported once you sign the Cold Outreach & Data Provenance Waiver, which puts the provenance representation, the consent basis and the responsibility for that program in writing. Scraped data is where we stop. You get that answer on the first call, not the fourth. ## Good reasons to take the call - Questions before starting the free sandbox - Planning a complex rollout or production workflow - Technical architecture: workspace, /v1, webhooks, and your Twilio setup - Enterprise terms: volume discounts, 3-year commitments, and security review Ready to build instead? The sandbox is free, and it takes about five minutes to open. --- # The per-call evidence file | Vocapable > One evidence record per attempt: what bound the dial, which checks ran, and which were waived under your signed attestation, with the signer named. URL: https://www.vocapable.com/evidence Product · The per-call evidence file # Every call leaves a file you can hand over One record per attempt, written while the call happens, whether it connected, reached a machine, or was stopped before it was ever placed. It carries the window and the caps that bound the dial in the contact’s local time, the consent record relied on, whether the AI-disclosure opener played and whether the carrier confirmed it, and which checks were waived under your signed attestation and by whom. A check that did not run never reads as a check that found nothing. ## What the file records Not a summary of the call. The state the gate read, the answer it returned, and the signatures behind anything it did not perform itself, in the order the snapshot writes them. - Where the dial stood The rules that bound this call, and the version they were at policy_jurisdiction · source_versions · timezone · local_time · quiet_hours Every set of rules that applied to this dial, each named with the version it was at when it applied, plus the callee’s own timezone, the local clock the gate computed, and the window that resulted. What the file settles is which of them bound the call, down to the pack and the version. - Whether anyone agreed to be called The consent records the decision actually stood on purpose · consent_required · consent_held · consent_expired · consent_covers_purpose · consent_record_ids One record has to carry both a sufficient class and the requested voice purpose; unrelated grants are never added together to reach a threshold, and a revocation row carrying an old record’s id is never accepted as a replacement grant. The ids on the snapshot are evidence, not permission: the records are re-read at dial time rather than trusted across an expiry or a withdrawal. - What ran One object per check that executed stages[] · stage · verdict · source · dataset_version · cache_age_days Normalization and suppression against your own internal do-not-call list are the floor stages, and they run for every tenant in every mode. Nothing reaches them, not a plan, not an API field, not a support tool, because the floor controls are not members of the waivable set at all, so “waive the disclosure opener” is not a value anyone can construct, express in JSON, or store. Making it unrepresentable is stronger than validating it away. - What was waived, and on whose signature One object per stage your attestation elected away compliance_elections[] · waived · attestation_ref · attested_by Each election names the stage, the attestation reference it stands on, and the officer who signed it. A waiver missing either half is refused when the object is constructed, so it cannot exist in memory, let alone reach a dial. The attestation is your own officer’s signature, not ours. - What happened on the call The re-check at launch, the opener, and the opt-out dial_time_recheck · disclosure · opt_out The gate re-runs as the channel is taken and stamps the result on the passing path too, so a dial that was re-verified at launch is distinguishable from one that only ever cleared queueing. The opener’s text lands with whether it was spoken and whether the carrier confirmed it. If the caller asked to stop, the file records when, what phrase matched, and what became of the suppression write. Spoken and confirmed are two different facts and the file keeps them apart. Spoken means the opener was handed to the transport. Confirmed means the carrier acknowledged the end of it, and only the second is evidence a person heard it. A call carrying no playback record at all is scored a failure rather than an unknown, because a dimension reporting “no data” would let the gap sit unnoticed forever. The proof, per call ## The file, in full A synthetic tenant and synthetic ids, on a Texas contact called in the middle of a Monday morning. This is the shape a real attempt carries, key for key. policy_snapshot att_01JB7Q4ZK9M2P6R3T5V7Y9AB2C Synthetic data An election is the shape a waiver is written in: the stage, the document it stands on, and the officer who signed it, because a stage nobody performed must never render as a stage that came back clean. The pack versions and the window are named with the versions they were at, so the file says which values bound the dial without saying anyone here judged them sufficient. Two millisecond playback offsets the real object can carry are left null here, since no timing figure belongs on a page like this one. Take it with you A file we authored to show the shape. It is not an export the product generated, and no customer data of any kind is in it. The part nobody else publishes ## Waived is a signature, not a silence Your file never goes quiet about a check nobody ran. Each one lands as a named election: the stage, the document it stands on, and the officer who signed it. A reviewer can follow that to a person, which is more than a green check has ever given anybody. The screening you run upstream is what those elections stand on. Every production customer signs one attestation naming that arrangement before the first production dial, and its reference and its signer are written into every attempt afterwards, so the file answers who screened this number and on which document. A stage that turns up with no election behind it is a defect on our side, never a waiver on yours, and the platform refuses to write a half-signed one: an election missing its reference or its signer is rejected when the object is built. Ran The check executed. Its answer sits in `stages[]` with the verdict, the source it came from, the dataset version behind it and how old that answer was. Waived Your attestation elected it away. The record carries the officer who signed that and the document it stands on, as its own object under its own key. Not applicable There was nothing to check. A caller-initiated conversation has no outbound window to evaluate, so the record carries a verdict of not applicable with “no outbound dial” as the reason, in its own words. Different shapes, under different keys. None of them can be mistaken for another by anyone reading the file, which is the entire point of writing it this way. ## The calls it refuses are evidence too A refusal is written with the same care as a connection, because the dial you did not place is the one you will be asked about. A number on your own suppression list is refused at the batch run and refused again when the channel is taken, and both refusals land on the record with the reason that produced them. The file keeps two vocabularies apart on purpose. One says the number could not be called at all. The other says the number was fine and this moment was not: outside the window, over the cap, or short of the consent class that purpose requires. Collapse those two and the file can no longer answer the first question anyone asks about a call: was this number callable at all, or just not right now? contact_attempt Synthetic data Suppression is re-checked at dial time, not only at import, and no plan, API field or support tool reaches it. ## How the file gets out The snapshot is a column on the attempt row, not a report you have to request. It leaves the platform three ways: the attempt rows and their snapshots as a CSV export, the screening results of a run as their own export kind (scrub_results), and the per-contact verdict live from the API before anything is dialed. Extractions are submitted as a job, and a completed job provides a signed download during its retention window. The pre-dial verdict is computed over that contact’s own record: your suppression list, the consent you hold, and the window and caps in force at that moment. It answers one question, whether this dial clears the gate right now, and it hands back the inputs it read alongside the answer, so a refusal arrives with the reason that produced it rather than as a number you have to go and investigate. ### The shape it arrives in The record sits on every attempt and the export carries it out with the rows intact. Ask for a matter and you get the attempt rows and their snapshots, which is the form your own counsel or your own tooling can load and search. ### What holds it still The audit tables are append-only at the database itself: no role holds UPDATE or DELETE on them, so a correction is a new row pointing at the row it corrects, and the application tier could not rewrite history even by bug. CSV export jobs let your team retrieve the selected attempt records and snapshots without changing that source history. The trust close ## A record of what happened, not an opinion about what was allowed The window, the caps, the consent standard and the recording-consent class are enforced on every dial, and the file names the ones that bound this call and dates the rule pack they came from. That distinction, what bound the call rather than whether it was enough, is what keeps the file usable in an answer years later. What your jurisdiction requires of your program is yours to determine with your own advisors. Our half is enforcement and evidence: the dials the platform can detect as ones that should not go out are stopped before the carrier is asked to originate anything, and every dial, placed or refused, stays reconstructible years later from the record alone. That is the claim still standing on the day somebody asks you to prove it. Every production dial rides your own Twilio subaccount, on your numbers, to the contacts on your own records. The choice of who to call is yours, and so is the screening behind it. Our half is enforcement and evidence. The rest of the posture, including the floor no plan disables, is on the Trust Center. ## What getting started actually looks like Your own Twilio account is the step that takes real time, and it is the reason your minutes cost what your carrier charges you. See pricing for the plans themselves. ### 1 · Build it in the sandbox The whole dial path runs against simulated telephony, every gate live and failing closed, and a record written the way production writes it, before any number rings anywhere. ### 2 · Sign, and say who screens A signed production order, your own Twilio subaccount, and the attestation that records your screening arrangement. Its reference and its signer ride into every evidence file from the first dial onward. ### 3 · Take the records out Pull the attempt rows and their snapshots by CSV export, or read them from the API. Your stack stays the system of record. ## Questions people ask first It is a column on the attempt row, written as the call happens, so nothing has to be assembled after the fact. It leaves through the exports API as CSV: submit an export job for the selected attempts, and a completed job provides a signed download during its retention window. What lands is the rows as the record wrote them, which is the form your own counsel or your own tooling can load and search. It names the check as waived, with the attestation reference it stands on and the officer who signed it. That is a different key and a structurally different object from a check that ran, so a stage nobody performed can never render as a stage that came back clean. A stage simply absent, with no election recorded, is a defect on our side rather than a waiver on yours. Your own compliance officer signs it, once, before the first production dial, and it names the screening arrangement you already run. From that point its reference and its signer are written onto your compliance profile and into every attempt the platform makes, so any call you are ever asked about can name the document behind it. Both halves are required in code: an election carrying a reference and no signer is rejected when the object is constructed. The database holds the other end, refusing any profile that carries waived stages with no attestation reference on it. Normalization, suppression against your own internal do-not-call list, and the dialability gate that applies the calling window, the caps and the consent standard in force. Suppression is re-checked at dial time and not only at import. Each check that runs is written into the stages array with its verdict and its source. It settles what happened: which values bound the dial, which checks ran, which your attestation covered and who signed for them, and how the call ended. What your jurisdiction requires of your program is yours to determine with your own advisors. The file’s job is to make every attempt, placed or refused, reconstructible from the record alone, years after anyone remembers the call. ## The record writes itself, on every call, on every plan Open a sandbox account and watch one get written, with nothing pointed at anyone. --- # The questions buyers ask first | Vocapable > Straight answers to what buyers ask first: what a call costs, whose carrier account it rides, what the agent does on the line, and what lands on your webhook. URL: https://www.vocapable.com/faq Questions # The questions buyers ask first What a call costs, whose account it rides, what the agent says, and what comes back when it hangs up. Every one is answered in exactly one place on this site, and this page is the index of where. The category questions are answered here. The rest are answered on the page that owns them, standing next to the mechanism they describe, so you get the answer and the thing it is about on one screen. ## Answered here Category questions rather than product-page details: the things a buyer types into a search box before they know which page of ours they want. - What does an AI calling platform actually cost? Every number is published on this site, so here is the whole shape of the bill rather than a range you have to call somebody for. Read the whole answer - Is AI cold calling legal in the US? The answer turns on your calling program rather than on the software, so here is exactly what Vocapable enforces on every dial and what stays yours to decide. Read the whole answer - Can an AI answer my business phone? Yes, and on the number you already publish: the receptionist answers on your own Twilio account, so the carrier relationship, the registrations and the caller-ID reputation stay yours. Read the whole answer - Can I run this without your dashboard? Yes. The product resources the portal uses are available through /v1, so your own stack can drive the workflow day to day. Signup, onboarding, and first-key setup run in the browser. Read the whole answer - How is my customer data protected? Your contacts, your calls and your evidence are fenced off inside the database itself, so one bug in application code cannot hand another customer your rows. Read the whole answer - What happens to my call recordings? They stay on the call record in your workspace, and they leave by CSV or API whenever you want them. Read the whole answer ## Answered on the page that owns them Each of these pages answers its own questions at the bottom, next to the rates, the records and the screens those answers are about. ### Pricing Published rates and terms, what the free sandbox includes, and what a production order covers. ### Receptionist The inbound product: the answering steps you compose, what claims a connection, and what an invoice carries. ### Compliance engine The four controls with no off state, the one gate every dial passes, and where a revocation lands. ### Evidence file Everything one call leaves behind, what each line of it means, and how you export the lot. ### Outcomes Every metric computed on your own campaigns, and the read behind each one. ### Telephony Whose carrier account the call rides on, what the invoice charges for, and what stays yours. ## The answer to most of these is: open one and look The sandbox is free, needs no card, and does not expire. Everything a question here describes is something you can go and check. --- # Can an AI answer my business phone? | Vocapable > Inbound answering on the number you already publish and the Twilio account you already own, following the steps you wrote down, in your order. URL: https://www.vocapable.com/faq/can-an-ai-answer-my-business-phone # Can an AI answer my business phone? Yes, and on the number you already publish: the receptionist answers on your own Twilio account, so the carrier relationship, the registrations and the caller-ID reputation stay yours. You compose what happens as an ordered chain of steps, one chain for your open hours and one for everything else, built from a keypress menu, ringing your own people, the AI agent, voicemail, a spoken message, or a forward to a number you configured. Open hours and holiday dates are in your own timezone. Only the AI step claims a reserved inbound connection. Menus, rings, forwards and spoken messages claim none, which is why a receptionist can publish an answering flow far larger than the concurrency it bought. Receptionist connections are reserved for inbound calls, separately from campaign capacity. When every connection you bought is busy, the caller advances to the next step you wrote down, usually your own voicemail. It answers in English and works from the steps you wrote. The call record keeps the disposition and available evidence—such as a transcript, recording, or voicemail—so your team has context for the next move. A Receptionist plan answers calls, and outbound campaigns run on an Engage, Workforce or Sales plan beside it. ## Where this is documented in full ### Receptionist The inbound product: the answering steps you compose, what claims a connection, and what an invoice carries. ## People who ask this also ask - What happens when every connection is busy? Receptionist - Can I run outbound campaigns on a Receptionist plan? Receptionist - Do I need my own Twilio account? Receptionist - What does an AI calling platform actually cost? Answered here ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # Can I run this without your dashboard? | Vocapable > Use /v1 for product resources, keep your stack as the system of record, and complete signup and onboarding in the browser. URL: https://www.vocapable.com/faq/can-i-run-this-without-your-dashboard # Can I run this without your dashboard? Yes. The product resources the portal uses are available through /v1, so your own stack can drive the workflow day to day. Signup, onboarding, and first-key setup run in the browser. The portal is a client of the data-plane API, so your code can work with contacts, lists, CSV imports and their row-level error reports, agents and immutable published versions, campaigns and their pacing, calls with available evidence, structured outcomes, consent and internal do-not-call records, exports, and analytics. Cursor pagination, RFC 9457 problem responses, scoped keys you rotate in place, and one signed webhook endpoint per tenant with delivery history and redrive controls support a headless workflow. Your stack stays the system of record, contacts can be created through the API, and configured events can be delivered to your endpoint. The compliance floor is identical either way: the same four controls, the same single gate, the same per-call file. Going live headless takes what going live always takes, which is a signed order, a payment method on file, and your own Twilio account connected and verified. Public developer documentation: https://docs.vocapable.com. ## Where this is documented in full ### Outcomes Every metric computed on your own campaigns, and the read behind each one. ### Pricing Published rates and terms, what the free sandbox includes, and what a production order covers. ### Telephony Whose carrier account the call rides on, what the invoice charges for, and what stays yours. ## People who ask this also ask - Can I get all of this without opening the dashboard? Outcomes - Do we have to upload our contact lists? Pricing - Does connecting my own carrier change which controls run? Telephony ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # How is my customer data protected? | Vocapable > Tenant isolation Postgres enforces underneath the application, encrypted credential storage, an append-only record, and evidence kept for five years. URL: https://www.vocapable.com/faq/how-is-my-customer-data-protected # How is my customer data protected? Your contacts, your calls and your evidence are fenced off inside the database itself, so one bug in application code cannot hand another customer your rows. Every contact, call, transcript and evidence row we hold for you carries your tenant id, and Postgres row-level security decides what a query may see before our code gets a vote. The service connects as a database role that owns no table and cannot bypass those policies, which is the whole point: the enforcement sits underneath the application rather than inside it. At startup the service inspects that role and refuses to serve if it is wrong, so a misconfigured deployment fails at boot instead of failing on your data. Nothing in the codebase reads your contacts, your calls or your evidence without a tenant bound to the transaction, and a query that reaches those tables with no tenant bound raises instead of returning rows. The failure mode is an error on our side, never another customer's contacts on yours. A test that gates every deploy proves it on a reused connection: bind one tenant, run its work, hand the same pooled connection to the next caller with nothing bound, and the database raises rather than serving the last tenant's rows. A pooled connection is exactly where this breaks quietly, so that is the case the test pins. Credentials you connect are held as envelope-encrypted blobs under AES-256-GCM, keyed from the secrets manager at deploy. The table stores ciphertext and an address, never a usable secret: the half that can decrypt cannot reach the database, the half that reads the database cannot decrypt, and CI fails a build that lets those two halves import each other. Our own provider secrets are injected at task launch and sit in no repository, no image and no CI variable. The record behind your calls is append-only in the database rather than by policy. No role holds UPDATE or DELETE on the audit table, so the application tier cannot rewrite history even by accident, and a correction is a new row pointing at the one it corrects. Per-call evidence, consent records and signed attestations are kept on a five-year retention class, because the file that answers a question about a call from three years ago has to still be there. A recording carries its own expiry, stamped 90 days out at the moment it is written, and a nightly job deletes the audio when that date lands. Your Data Processing Addendum governs retention, and it is published in full before you sign it. Access is the last layer. An API key is stored as a SHA-256 hash with a short display prefix, shown in full exactly once, scoped per route and rotatable in place; test and live credentials are separated at the credential level, and a test credential cannot originate a live dial. Passwords are argon2id with a deployment pepper, one comparison revokes every session on every device, and each staff action on the admin surface writes its own audit row. Send a security questionnaire to support@capstralabs.com and you get written answers about these controls from the people who built them. ## Where this is documented in full ### Compliance engine The four controls with no off state, the one gate every dial passes, and where a revocation lands. ### Evidence file Everything one call leaves behind, what each line of it means, and how you export the lot. ## People who ask this also ask - Can any plan, flag, or support ticket switch the four controls off? Compliance engine - Is the evidence file something I download from a dashboard? Evidence file - What happens to my call recordings? Answered here - Can I run this without your dashboard? Answered here ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # Is AI cold calling legal in the US? | Vocapable > What Vocapable enforces on every dial, the artificial-voice rule the FCC set in 2024, and the record every call leaves behind for you. URL: https://www.vocapable.com/faq/is-ai-cold-calling-legal-in-the-us # Is AI cold calling legal in the US? The answer turns on your calling program rather than on the software, so here is exactly what Vocapable enforces on every dial and what stays yours to decide. What decides it is who you called, what consent you held for that number, where the person was, and when their phone rang. Vocapable enforces the policy in force on every dial and writes it into the record. That is an enforcement mechanism, not a reading of the law. One thing about the technology is settled enough to build on: the FCC has treated a call made with an AI-generated voice as an artificial-voice call under the TCPA since February 2024, which is why the AI disclosure here is compiled by the policy engine, played before the model produces its first token, and reachable by no plan, flag, API field or support tool. The rest is mechanism and record. Consent is a typed class ranked against the standard in force, calling windows and frequency caps are evaluated in the contact’s local time, one write path carries a revocation across every channel it covers, and every call leaves an exportable file, including the calls the platform refused to place. That file names the checks that ran, the values that bound the dial and the signatures behind them, and it exports by CSV or API whenever a reviewer asks for it. No control on this platform settles whether a call was lawful, and Vocapable does not claim otherwise. The agreement says it in the same words, and the waiver a purchased, rented or appended list runs under records who carries that responsibility. ## Where this is documented in full ### Evidence file Everything one call leaves behind, what each line of it means, and how you export the lot. ### Compliance engine The four controls with no off state, the one gate every dial passes, and where a revocation lands. ## People who ask this also ask - Which checks does the platform itself perform on a dial? Evidence file - What can these controls not do? Compliance engine - Does an evidence file settle whether a call was lawful? Evidence file - What happens to my call recordings? Answered here ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # What does an AI calling platform actually cost? | Vocapable > The whole shape of the bill: a monthly plan, a per-call fee, AI usage by the minute, and why the carrier minutes stay on your own Twilio bill. URL: https://www.vocapable.com/faq/what-does-an-ai-calling-platform-cost # What does an AI calling platform actually cost? Every number is published on this site, so here is the whole shape of the bill rather than a range you have to call somebody for. You pay a monthly plan price that carries your calling capacity, $0.005 per call attempt on the calls your account places, and AI usage by the minute: $0.13 per minute month to month, or $0.09 per minute on a signed 3-year commitment, stepping down with total monthly spend to a published floor of $0.07 per minute. Outbound plans start at $349 a month. Inbound answering starts at $250 a month, and a Receptionist invoice is that plan price plus any extra connections you buy, because it answers calls rather than placing them. The carrier line is the one most comparisons leave out. Production calls leave your own Twilio account at the rate you already negotiated, and Twilio bills you for those minutes directly, so your Twilio rate is the rate you pay. The per-minute figure above prices the AI session, which is why it reads the same whoever holds the carrier account. Every figure is on the pricing page, implementation is $0, and the sandbox in front of all of it is free. Price your own campaign shape against these same rates before you talk to anybody. ## Where this is documented in full ### Telephony Whose carrier account the call rides on, what the invoice charges for, and what stays yours. ### Pricing Published rates and terms, what the free sandbox includes, and what a production order covers. ### Compliance engine The four controls with no off state, the one gate every dial passes, and where a revocation lands. ## People who ask this also ask - What does a Vocapable invoice charge for the calling? Telephony - Can these prices change? Pricing - Is the compliance floor a paid add-on? Compliance engine ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # What happens to my call recordings? | Vocapable > Where a recording lives, what the per-call evidence file records beside it, how both leave by CSV or API, and which document governs retention. URL: https://www.vocapable.com/faq/what-happens-to-my-call-recordings # What happens to my call recordings? They stay on the call record in your workspace, and they leave by CSV or API whenever you want them. A recording sits on the contact’s activity timeline beside that call’s transcript, its disposition, the consent basis relied on and the internal do-not-call state. One place, rather than a recording in one system and the reason for the call in another. The same call’s compliance evidence is written separately as a per-call file recording which checks ran and which values bound the dial. Recording consent is applied as a control rather than an opinion. The recording-consent standard in force is enforced at the gate on every dial and recorded on the attempt, alongside the calling window and the caps that bound it. Retention, deletion, and what happens to the audio when an agreement ends are written into the Data Processing Addendum and the Privacy Policy, and both are published in full before anybody signs. Vocapable serves US calling, and its privacy machinery is built on CCPA and CPRA. Export is the part worth testing early. Pull a call, its transcript and its evidence file through the exports API in your first week, and you know exactly what you can hand a reviewer five years from now. ## Where this is documented in full ### Evidence file Everything one call leaves behind, what each line of it means, and how you export the lot. ### Compliance engine The four controls with no off state, the one gate every dial passes, and where a revocation lands. ## People who ask this also ask - Is the evidence file something I download from a dashboard? Evidence file - What does the file say about a check that nobody ran? Evidence file - When does an opt-out take effect? Compliance engine - Is AI cold calling legal in the US? Answered here - How is my customer data protected? Answered here ## Check it yourself, before anybody signs anything The sandbox is free and needs no card. Build the agent, run the whole dial path with every compliance gate live, and read the file it writes. --- # Start free in the sandbox | Vocapable > Start free in the sandbox. Build your agent and hear the simulated demo calls. You pay when you go live. URL: https://www.vocapable.com/get-started Get started # Start free. Go live when you are ready. A free sandbox account gives your team the whole builder, with your rubric, your material, and your voice in it. Build the workflow, test it, then choose a plan, accept the production order, add a payment method, and connect your verified Twilio account in the portal. You pay when you activate production, and not a day before. Create your account, verify your email, answer a few setup questions, and your sandbox opens. About five minutes, no credit card. Already have an account? Sign in ## How it works: three steps, no surprises Today · Free ### Start in the sandbox Create an account, verify your email, answer a few setup questions, and your sandbox opens. Free, no credit card, and your team gets the full builder: direct no-tool drafts, knowledge packs, your rubric, your voice. Simulated test calls are free and never metered: the full dial path, every compliance gate live and failing closed, every refusal and disposition written the way production writes them. When you’re ready ### Choose your plan Choose your workload class and term, accept the Production Services Order with its published rates, and add a payment method. The portal keeps the next production-readiness step visible, so the configuration you tested carries forward. Go live ### Activate on your carrier Connect and verify a Twilio subaccount you own in the portal: your numbers, registrations, negotiated rates, and caller-ID reputation stay with you. Ready workspaces activate production there. If a carrier or compliance decision needs attention, the same readiness context routes to review. ## What the free sandbox includes The whole product runs here. Your real use case, your real material, your rubric and your voice. Then run it: the dial path executes end to end, every compliance gate live and failing closed, every refusal evidenced and every disposition written the way production writes it. Production dialing begins when a ready workspace completes activation in the portal. - Full agent builder: personas, objectives, guardrails, knowledge packs - Direct no-tool conversation drafts: qualification, staffing, and feedback - Workspace or headless: run it from the dashboards, or drive it entirely from your own stack via the API - Unlimited simulated test calls, free - Immutable agent versions behind a publish gate: what you tested is exactly what dials - The complete compliance floor, included at $0 on every plan - No credit card: you pay when you activate production, and not a day before Why a production order? The order keeps your selected rates and commitments in one place. You accept it in the portal as part of production readiness, with every term in writing before a single customer hears your agent. You bring the list and choose who on it gets called, and that choice stays yours. Ours is enforcement, on every attempt: the calling window in the contact’s local time, the frequency caps, the consent standard and the recording-consent class, with the values that bound each call written onto the call. Sandbox first means you activate knowing exactly what you built, not what a slide deck promised. Rates and terms are on the pricing page. Prefer to talk first? Talk to sales. --- # Glossary of AI calling and TCPA terms | Vocapable > Plain-English definitions of the terms in an AI phone-outreach contract, so you can read your own agreement and your own invoice and know what every line means. URL: https://www.vocapable.com/glossary Reference · Glossary # AI calling, defined term by term Read your own agreement and your own invoice with confidence. Each entry gives you one term in plain English, what it costs you or commits you to, and the page where you can watch the same thing run in the product. How to use it ## Written for the person who signs the contract Every term here lands somewhere that costs you money or hands you a duty: a line on your invoice, a clause in the agreement, a field on a contact record. So each entry answers the four questions a buyer actually has. What is this. Where does it start and where does it stop. Who owns it, you or us. And what does the record show six months later, when somebody asks. Then it hands you the next step. Every definition links the page where the same subject is a mechanism you can read, question, and hold us to, so you can check a word against the product in one click instead of taking it from a definition. Money stays on the page that publishes it. A rate typed into a definition drifts away from the rate you are billed, so an entry that touches your bill links therate card rather than copying a figure out of it. These pages define words. What your own calling program has to do where you dial is yours to determine with your own advisors. Thecompliance engine is what the software enforces on every call, theper-call evidence file is what it writes down, and theTrust Center holds the agreements you would be signing. ## Every definition - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening Grouped by where each one leads ## Every definition owes you a next step Each entry below names the page where the same subject is described as a mechanism, with the price it carries and the record it leaves. Read the word here, then check it there. ### Pricing Published plans and usage rates. The sandbox is free to start. #### Call attempt vs conversation Two things the platform counts separately, and bills separately. #### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. ### A CRM built for one job: the phone Contacts, calls, caller requests, and outcomes in one workspace. #### Established business relationship (EBR) A calling basis resting on a relationship already in your records. #### Prior express written consent (PEWC) The strictest consent class, its disclosure text kept verbatim. ### AI receptionist for your inbound calls Your inbound calls, answered on your own Twilio, in your order. #### Concurrent inbound connection Inbound answering capacity, reserved so outbound cannot eat it. ### Compliance enforced in code Four controls no plan disables, and a file for every call. #### Internal do-not-call list Your own suppression list, read on every call the platform places. #### Revoke-all revocation One person asking you to stop, applied everywhere in one write. #### TCPA The statute governing calls placed with an artificial voice. #### Upstream list screening The screening you run on your own list, recorded on every call. ### The per-call evidence file Every call leaves a file you can hand to a reviewer later. #### Per-call evidence file What every call leaves behind, including the ones we refused. ### Your Twilio, your numbers, your carrier bill Your carrier account, your numbers, your caller-ID reputation. #### BYO telephony Your Twilio subaccount, your numbers, your carrier bill. Always. ### Outcomes: the metrics we report Named metrics, computed on your campaigns. Nothing invented. #### Answering-machine detection (AMD) What the platform does when a machine picks up, not a person. #### Speed-to-lead The time from a lead arriving to the first attempt leaving. ## See one of these words in your own account Open a free sandbox account and read what one of these terms looks like once the software writes it down: the checks that ran on a call, the values that bound the attempt, and the reason it went out. --- # Answering-machine detection (AMD) | Vocapable > Carrier-side classification of who or what answered a dial, and what the dialer does next when the answer turns out to be a machine rather than a person. URL: https://www.vocapable.com/glossary/answering-machine-detection Glossary # Answering-machine detection (AMD) What the platform does when a machine picks up, not a person. Published August 18, 2026 Answering-machine detection, usually shortened to AMD, is carrier-side classification of who or what picked up, requested on every dial so that the agent is not holding a conversation with a greeting. On a campaign that did not elect a voicemail message — the default — a machine verdict hangs the call up immediately. Nothing is spoken to the machine, no conversation slot is claimed, and the dialer then takes an explicit retry-or-dispose branch that is written onto the attempt itself. The outcome is recorded rather than disposed silently, which is the bar every disposition on this platform is held to. Detection is imperfect in both directions, and both directions are treated as compliance events rather than as quality noise. A live person misclassified as a machine hears silence and a hang-up, which from their side is indistinguishable from an abandoned call, so it counts against the same abandonment budget the pacer is governed by, and the classifier’s accuracy is sampled rather than assumed. Where a campaign has elected to leave an artificial-voice voicemail message under an executed Cold Outreach and Data Provenance Waiver, that message is itself a regulated call. It carries the AI-disclosure opener composed from the same pinned agent version a live conversation discloses from, the internal suppression list is re-read at the instant it is left, and no contact receives more than one such message per campaign. Vocapable records the text it dispatched, that text’s hash and the waiver in force at that moment. What every attempt leaves on the record afterwards is on theoutcomes page. Where this goes next ## Outcomes: the metrics we report Named metrics, computed on your campaigns. Nothing invented. ## Related definitions ### Call attempt vs conversation Two things the platform counts separately, and bills separately. ### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. ### Per-call evidence file What every call leaves behind, including the ones we refused. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # BYO telephony | Vocapable > Bring-your-own telephony: every production call leaves on your own Twilio subaccount, from numbers in your name. Vocapable originates no production dial. URL: https://www.vocapable.com/glossary/byo-telephony Glossary # BYO telephony Your Twilio subaccount, your numbers, your carrier bill. Always. Published August 18, 2026 Bring-your-own telephony, usually shortened to BYO: every production call Vocapable places for you leaves on your own Twilio subaccount, from numbers registered in your name, under your own registrations. The carrier account, the negotiated per-minute carrier rates, the caller-ID reputation you spent years building and the carrier invoice all stay yours. Every production dial originates on your account, so the entity a carrier or a regulator sees as the caller is the entity that actually is the caller, and a year of answer rates compounds on numbers sitting in your own name. The money follows the ownership. Your carrier bills you for the carriage directly, at the rates you negotiated, so a Vocapable invoice carries no carrier line: no minutes bought from us, no number rental, nothing to reconcile against a bill we never saw. Win a better rate from your carrier and the saving lands in your pocket that month. It is the structure rather than a promotion, so it does not expire when a term does. Platform-owned numbers carry only paid test calls to numbers you have proved you control, and nothing else. Vocapable runs on Twilio; no partnership or endorsement is implied. What connecting your own account involves is on thetelephony page. Where this goes next ## Your Twilio, your numbers, your carrier bill Your carrier account, your numbers, your caller-ID reputation. ## Related definitions ### Call attempt vs conversation Two things the platform counts separately, and bills separately. ### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. ### Upstream list screening The screening you run on your own list, recorded on every call. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Call attempt vs conversation | Vocapable > An attempt is one dial handed to your carrier; a conversation is what happens after a person answers. The outbound bill is built from both, counted separately. URL: https://www.vocapable.com/glossary/call-attempt-vs-conversation Glossary # Call attempt vs conversation Two things the platform counts separately, and bills separately. Published August 18, 2026 Two different things the platform counts separately, and the distinction is most of how outbound is priced. An attempt is one dial handed to your carrier — connected or not, answered by a person, by a machine, or by nobody. Every outbound plan pays a published per-attempt fee for it. An attempt the platform refused before a dial ever existed is charged nothing at all, which is worth saying plainly: a compliance refusal costs you nothing except the contact you should not have called. A conversation is what happens once a person answers. It claims a concurrent session, and its minutes bill as AI usage per minute — the per-minute rate prices the AI, not the telephony, which is why it applies on a bring-your-own-carrier account exactly as it does anywhere else, while your carrier bills you separately for the carriage. Reaching the same person on a third attempt is still one conversation, which is why an estimate’s attempt count follows the retry curve while its conversation count can never exceed the size of your list. The estimator runs on your own numbers: the answer rate you already know for your list, your retry ceiling, your average call length. It leaves the volume discount out on purpose, so the figure you compute can come in above the invoice and never below it. The per-attempt fee, the per-minute rates and the volume-discount floor are all on thepricing page rather than typed into this definition. Where this goes next ## Pricing Published plans and usage rates. The sandbox is free to start. ## Related definitions ### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. ### Answering-machine detection (AMD) What the platform does when a machine picks up, not a person. ### BYO telephony Your Twilio subaccount, your numbers, your carrier bill. Always. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Concurrent inbound connection | Vocapable > The inbound twin of a concurrent session: how many callers the AI receptionist can talk to at once, reserved against your own outbound campaigns. URL: https://www.vocapable.com/glossary/concurrent-inbound-connection Glossary # Concurrent inbound connection Inbound answering capacity, reserved so outbound cannot eat it. Published August 18, 2026 A concurrent inbound connection is the inbound twin of a concurrent session, and a separate entitlement rather than the same pool: it is how many callers your AI receptionist can be talking to at the same moment. The two sides of the product are bought separately, which is what makes the answering capacity dependable: your own dialer at full pace cannot take a connection you bought to answer with. A purchased connection is reserved rather than shared. It is subtracted from what your own outbound campaigns may claim, at the tenant level and again at the fleet level, so a campaign you launched cannot starve the inbound capacity you are paying for. The reservation runs against your own traffic, which is the only traffic it can honestly run against. Only the step where the AI actually answers claims one. A menu, a ring through to your own team, a forward, or a static spoken message runs on the carrier’s own leg: no conversation slot, no model, no meter. Under BYO those seconds are billed to you by your own carrier, exactly as every other second on your line is. When every connection you bought is busy, that is an ordinary operating state of a product sold by the connection rather than an error — the caller advances to whatever you configured next, usually voicemail. The plan price, the included count and the price of an additional connection are published on the pricing page; what the inbound product does with a call once it answers is on thereceptionist page. Where this goes next ## AI receptionist for your inbound calls Your inbound calls, answered on your own Twilio, in your order. ## Related definitions ### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. ### BYO telephony Your Twilio subaccount, your numbers, your carrier bill. Always. ### Answering-machine detection (AMD) What the platform does when a machine picks up, not a person. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Concurrent session | Vocapable > The unit outbound plans are sold in: one live AI conversation running at the same time as another, claimed when a person answers and released at hangup. URL: https://www.vocapable.com/glossary/concurrent-session Glossary # Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. Published August 18, 2026 A concurrent session is one live AI conversation running at the same time as another: speech in, model, speech out — pipeline capacity on the GPU plane, not a carrier line. It is the unit outbound plans are sold in, and reading it as a phone line is the most common mistake a buyer makes on the pricing page. A call claims one from the moment a person answers until the moment they hang up, and it is claimed by the voice runtime rather than by the dialer, because the capacity is only needed once somebody actually picks up. A dial that rings out, reaches a machine on a campaign with no message elected, or is refused before it ever leaves claims nothing at all. Each outbound plan’s monthly price includes a set number of them, and additional sessions are sold monthly in any quantity, bought and changed from the billing page rather than negotiated with a salesperson. The included counts and the price of an extra session live on the pricing page and are deliberately not retyped here: a rate copied into a definition is a rate that drifts away from the one you are billed. Inbound callbacks claim from the same pool. A burst of people ringing you back narrows the headroom your outbound campaigns are pacing against, automatically and with nothing for you to configure. Where this goes next ## Pricing Published plans and usage rates. The sandbox is free to start. ## Related definitions ### Concurrent inbound connection Inbound answering capacity, reserved so outbound cannot eat it. ### Call attempt vs conversation Two things the platform counts separately, and bills separately. ### Answering-machine detection (AMD) What the platform does when a machine picks up, not a person. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Established business relationship (EBR) | Vocapable > A calling basis resting on a purchase or an inquiry already recorded in your own systems, on the windows the Telemarketing Sales Rule sets for each one. URL: https://www.vocapable.com/glossary/established-business-relationship Glossary # Established business relationship (EBR) A calling basis resting on a relationship already in your records. Published August 18, 2026 An established business relationship, usually shortened to EBR, is a calling basis resting on a relationship the called party already has with you rather than on a signature they gave you. The Telemarketing Sales Rule sets the windows: a purchase, rental, lease or other financial transaction within the 540 days immediately before the call, or an inquiry or application about your goods or services within the 90 days immediately before it. It is the basis most feedback and service calling genuinely runs on, and it is the sturdier one in practice, because the transaction sits in your own records with a date on it — a basis you can actually produce if someone asks you for it. Vocapable maps a campaign’s purpose class to the consent scope that purpose requires, and where the two disagree the attempt is suppressed as a scope mismatch rather than dialed. Consent to service notifications does not authorize a sales campaign, and the suppression names that as its reason instead of disposing the attempt quietly, so the gap shows up on a report rather than in a demand letter. Whether the relationship qualifies where you dial is yours and your counsel’s to decide. What the platform does is record the basis you declared, read it before every attempt, and refuse the mismatch. How that basis is stored against a contact is on thecontact record page. ## Sources Each one carries the date we last read it, so you can open the primary text yourself and see how current the reading above is. - 16 CFR § 310.2(q) — Telemarketing Sales Rule definitions · read 2026-08-18 Where this goes next ## A CRM built for one job: the phone Contacts, calls, caller requests, and outcomes in one workspace. ## Related definitions ### Prior express written consent (PEWC) The strictest consent class, its disclosure text kept verbatim. ### TCPA The statute governing calls placed with an artificial voice. ### Internal do-not-call list Your own suppression list, read on every call the platform places. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Internal do-not-call list | Vocapable > Your own per-tenant suppression list, applied to the batch before a campaign runs and read live again before each dial, with rows that are never deleted. URL: https://www.vocapable.com/glossary/internal-do-not-call-list Glossary # Internal do-not-call list Your own suppression list, read on every call the platform places. Published August 18, 2026 Your own suppression list, held per tenant, and the one list the platform reads on every call it places. Every opt-out the agent hears writes a durable suppression row before the call closes, and the same list is applied to the batch before a campaign runs and read live again in the instant before each dial — so an opt-out captured moments ago on another campaign suppresses the very next attempt rather than the next batch. It is a floor control in the strict sense. No plan tier, feature flag, API field, support tool or compliance-mode setting switches it off, and if the live read is unavailable, dialing halts rather than proceeding without it. There is no configuration surface anywhere in the product that can name this control, which is why there is nothing to turn off — the absence is structural rather than a policy somebody is expected to keep. Its rows are permanent. They survive erasure requests, churn and non-payment, because deleting one would cause exactly the recontact the person asked you to prevent. The data processing addendum states the carve-out in those terms: suppression, not amnesia. The payoff is the one your risk register cares about. A person who asked you to stop is not called again by any campaign in your workspace, and you can produce the row that stopped it. How the arrangement you run on your own list is recorded beside it is covered under upstream list screening, and what the software performs on each dial is on thecompliance engine page. Where this goes next ## Compliance enforced in code Four controls no plan disables, and a file for every call. ## Related definitions ### Revoke-all revocation One person asking you to stop, applied everywhere in one write. ### Upstream list screening The screening you run on your own list, recorded on every call. ### Per-call evidence file What every call leaves behind, including the ones we refused. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Per-call evidence file | Vocapable > The exportable record every call leaves behind, including calls the platform refused to place: which checks ran, which were waived, and on whose signature. URL: https://www.vocapable.com/glossary/per-call-evidence-file Glossary # Per-call evidence file What every call leaves behind, including the ones we refused. Published August 18, 2026 The exportable record every call leaves behind, including the calls the platform refused to place. It records which checks ran, which were waived, and on whose signature. A check that did not run is written as a named election carrying its attestation reference and its signer — never as an absent row, and never as a check that found nothing. That last distinction is the whole design: an empty result and a stage nobody performed look identical in most systems, and they are the difference between a record that helps you and a record that hurts you. It carries the calling-policy values that bound the attempt together with the provenance of each one, so the file says who set the window and the caps and when they set them. A value the platform applied never reads as a value the platform vouched for. It also holds the consent record relied on with its content hash, the quiet-hour and attempt-cap evaluations computed against the callee’s own local clock, the pinned agent version and the model identifiers, and, where the call connected, confirmation that the AI-disclosure opener actually played. The conversation itself is kept as turns and a transcript. Every attempt produces exactly one of these, connected or not, and a call with no attempt behind it carries its snapshot on the call record instead — one home per call, never both. They live on an append-only store that no application role can update or delete: a correction is a new row referencing the corrected one, and they are excluded from tenant-initiated deletion because they are the platform’s own defense record rather than tenant content. Given a phone number and a date range — the exact shape of a demand letter — the platform assembles the bundle as an export job. A worked example is on theevidence page. Where this goes next ## The per-call evidence file Every call leaves a file you can hand to a reviewer later. ## Related definitions ### Upstream list screening The screening you run on your own list, recorded on every call. ### Internal do-not-call list Your own suppression list, read on every call the platform places. ### Prior express written consent (PEWC) The strictest consent class, its disclosure text kept verbatim. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Prior express written consent (PEWC) | Vocapable > The strictest consent class Vocapable records on a contact, held as an append-only row carrying its scope, its proof artifact and the verbatim disclosure text. URL: https://www.vocapable.com/glossary/prior-express-written-consent Glossary # Prior express written consent (PEWC) The strictest consent class, its disclosure text kept verbatim. Published August 18, 2026 Prior express written consent, usually shortened to PEWC, is the strictest of the consent classes Vocapable records on a contact: a signed writing, carrying the disclosure text that was actually shown to the person, agreeing to receive AI-voice or prerecorded marketing calls at a specific number. It is what federal rules require for artificial or prerecorded-voice telemarketing to residential lines and cell phones. In the product it is an append-only record carrying its scope — which purposes, which campaign categories, which channels — plus its source, its proof artifact and that artifact’s content hash, the capture time, and the verbatim disclosure text. A record offered as this class without that verbatim text is rejected before either row is written, so the class cannot exist in the database in a shape that could not be produced later. The class carries an expiry, and a tenant may configure a shorter life for it but never a longer one. Stricter-only is the direction every consent setting on the platform is allowed to move in, which is why a support conversation cannot loosen one. The declaration and its proof reference are yours. The platform reads them before every attempt, enforces what you declared, and carries both into the evidence file that attempt leaves behind, so the record is already assembled on the day somebody asks for it. Where consent lives, how it is captured and what a contact record carries alongside it are on the contact record page. ## Sources Each one carries the date we last read it, so you can open the primary text yourself and see how current the reading above is. - 47 CFR § 64.1200 — Delivery restrictions · read 2026-08-18 Where this goes next ## A CRM built for one job: the phone Contacts, calls, caller requests, and outcomes in one workspace. ## Related definitions ### TCPA The statute governing calls placed with an artificial voice. ### Established business relationship (EBR) A calling basis resting on a relationship already in your records. ### Per-call evidence file What every call leaves behind, including the ones we refused. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Revoke-all revocation | Vocapable > A revocation that converges on one database transaction, cascading suppression across every active campaign for the tenant and cancelling pending dials. URL: https://www.vocapable.com/glossary/revoke-all-revocation Glossary # Revoke-all revocation One person asking you to stop, applied everywhere in one write. Published August 18, 2026 One person asking you to stop, applied everywhere at once. A revocation spoken on the call, entered on the keypad, written in the dashboard or sent to the API converges on a single database transaction that appends the revocation record, writes the internal suppression row, withdraws the consent standing behind the number and sets the dial-time fast path together. Everything downstream follows from that one write. Every active campaign for the tenant stops calling that number, pending dials are cancelled rather than drained, and the next dial-time read finds the suppression already in place. A narrowed revocation is the dangerous failure here, so the write over-suppresses rather than under-suppresses by design: a person who asked you to stop about one thing is not called about another while somebody works out whether they meant it. Federal rules give a caller a reasonable time not to exceed ten business days to honor a revocation made in any reasonable manner, and they treat a revocation received by any such method as definitive rather than as scoped to the campaign it arrived on. Vocapable does it in the same write. The ten business days is the bar the rules set; the same transaction is a property of our write path, and it leaves your team nothing to schedule, nothing to chase and nothing to reconcile afterwards. The one revocation write path, and what it leaves on the record afterwards, are on thecompliance engine page. ## Sources Each one carries the date we last read it, so you can open the primary text yourself and see how current the reading above is. - 47 CFR § 64.1200(a)(10) — revocation of prior express consent · read 2026-08-18 Where this goes next ## Compliance enforced in code Four controls no plan disables, and a file for every call. ## Related definitions ### Internal do-not-call list Your own suppression list, read on every call the platform places. ### TCPA The statute governing calls placed with an artificial voice. ### Per-call evidence file What every call leaves behind, including the ones we refused. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Speed-to-lead | Vocapable > Speed-to-lead is the gap between a lead arriving in your system and the first call attempt leaving for it. Vocapable reports it as lead-to-call time. URL: https://www.vocapable.com/glossary/speed-to-lead Glossary # Speed-to-lead The time from a lead arriving to the first attempt leaving. Published August 18, 2026 Speed-to-lead is the time between a lead arriving in your system and the first call attempt leaving for that lead. Vocapable reports it as lead-to-call time, one of the named metrics the product computes off its own call ledgers rather than a figure anyone estimates on your behalf. It is the whole argument for putting an AI agent on inbound leads: the same list, the same qualification rubric and the same carrier account, worked while the lead is still warm instead of whenever a rep next reaches the queue. It keeps working after your team goes home, and the figure you manage against is your own, read off your own ledgers. The metrics reported alongside it are contact rate, qualification rate, refusals by reason, call-to-handoff time and cost per outcome — each computed from the platform’s own attempt and conversation records, each traceable back to the rows it was derived from. Every one of those figures describes the traffic you ran, on your own list, through your own carrier account. The reporting set is set out in full on theoutcomes page. Where this goes next ## Outcomes: the metrics we report Named metrics, computed on your campaigns. Nothing invented. ## Related definitions ### Call attempt vs conversation Two things the platform counts separately, and bills separately. ### Concurrent session The unit outbound plans are sold in: one live AI conversation at a time. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # TCPA | Vocapable > The federal statute governing calls placed with an artificial or prerecorded voice, and why this platform fails closed rather than dialing on an open question. URL: https://www.vocapable.com/glossary/tcpa Glossary # TCPA The statute governing calls placed with an artificial voice. Published August 18, 2026 TCPA is the federal Telephone Consumer Protection Act, the statute that governs calls placed with an artificial or prerecorded voice. An AI voice is an artificial voice for this purpose, so telemarketing calls to residential lines and cell phones require prior express written consent — and a purchased list never carries it, because no amount of screening rescues a call that required consent the caller never had. That is the premise this platform was built on rather than a disclaimer bolted onto it. When a check cannot complete, the dial does not happen: the attempt is disposed with its reason recorded rather than proceeding on an unanswered question, and there is no fail-open path anywhere in the authorization sequence. The controls a customer cannot switch off exist because the ones that can be switched off are worth nothing in a dispute. The division of duty is the part buyers most often get wrong, so it is stated plainly. You and your counsel decide what the law requires where you dial. Vocapable enforces the calling policy you set, records the values that bound every attempt and the name of whoever set them, and proves afterwards exactly what it applied. What the software does on a dial — the floor controls, the single authorization gate, the consent taxonomy and the record each attempt leaves — is set out on thecompliance engine page. ## Sources Each one carries the date we last read it, so you can open the primary text yourself and see how current the reading above is. - 47 CFR § 64.1200 — Delivery restrictions · read 2026-08-18 Where this goes next ## Compliance enforced in code Four controls no plan disables, and a file for every call. ## Related definitions ### Prior express written consent (PEWC) The strictest consent class, its disclosure text kept verbatim. ### Established business relationship (EBR) A calling basis resting on a relationship already in your records. ### Revoke-all revocation One person asking you to stop, applied everywhere in one write. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Upstream list screening | Vocapable > The screening you run on your own list before a dial, recorded on every call as a signed election naming the arrangement, its signer and its date. URL: https://www.vocapable.com/glossary/upstream-list-screening Glossary # Upstream list screening The screening you run on your own list, recorded on every call. Published August 18, 2026 · Updated August 20, 2026 Upstream list screening is the work you do on your own contact list before any of it reaches a dialer. You pick the sources, you run the queries under your own account, and the results stay in your hands. Vocapable’s job starts where yours stops. Every production account signs one attestation naming the arrangement it runs, the officer who signed it and the date. From that moment each stage the attestation covers rides into every call’s evidence file as your election, carrying that reference and that signer, with nothing for anyone to remember to attach. That gives you a record an auditor can read months later without asking anybody a question. A stage carried as an election reads as an election forever: checked and clear, checked and suppressed, and elected away under signature are three different rows, and the file keeps them apart. It is the difference between a record that helps you and one that hurts you. The same mechanism covers your paperwork gap. An account with no attestation on file screens its contacts undialable, with the missing document named, so a dial waits for a signature rather than going out on an unanswered question. On every dial, whatever your arrangement covers, the platform runs phone normalization, internal-DNC suppression at the batch run and again at dial time, and the dialability gate, which refuses rather than proceeding when it cannot answer. The controls that run on every call, and the record each one leaves, are on thecompliance engine page. Where this goes next ## Compliance enforced in code Four controls no plan disables, and a file for every call. ## Related definitions ### Internal do-not-call list Your own suppression list, read on every call the platform places. ### Per-call evidence file What every call leaves behind, including the ones we refused. ### BYO telephony Your Twilio subaccount, your numbers, your carrier bill. Always. This page defines a word. What your own calling program has to do where you dial is yours to determine with your own advisors. What the software enforces on every call is on thecompliance engine page, and every agreement you would sign is published at/legal. ## More definitions - Answering-machine detection (AMD) - BYO telephony - Call attempt vs conversation - Concurrent inbound connection - Concurrent session - Established business relationship (EBR) - Internal do-not-call list - Per-call evidence file - Prior express written consent (PEWC) - Revoke-all revocation - Speed-to-lead - TCPA - Upstream list screening - All definitions Every entry gives you one term in plain English, and the page where you can watch the same thing run in the product. --- # Vocapable: AI voice agents that turn calls into work > AI voice agents for outreach, reception, workforce, and feedback—turning calls into structured outcomes in a shared workspace. URL: https://www.vocapable.com/ # Make business calls useful. Qualify leads, coordinate your workforce, collect feedback, and give inbound callers a useful first answer. Start with the workflow your team needs today. Add reception, workforce outreach and sales qualification in the same workspace as your operation grows. Build around your business knowledge and questions, then keep call records, captured details and follow-up together. Your Twilio account keeps carrier control with you; Vocapable keeps call records, policy evidence, and next steps close to the work. Qualified leads Product UI · synthetic data Your list, end to end ## Watch a list become pipeline Your list goes in. The campaign works its eligible contacts within the calling window and caps. Completed conversations return as qualified outcomes your team can review, route, and act on. How a campaign runs. The list is yours, captured or acquired, with provenance attested at import and cold outreach signed for under the waiver you accept. You choose who on it gets called. Imported or delivered by API, at an illustrative size of 200,000, it passes three pre-dial checks before any number is called: normalization, suppression against your own internal do-not-call list, and the calling window and frequency caps. Refusals are recorded. Campaign attempts use your own Twilio subaccount and numbers. Connected conversations open with an AI disclosure and can be qualified against your rubric. Completed-call outcomes sort three ways: interested and opted in, with qualification and next-step details available to your workspace and systems; not now, with the reason documented and a retry inside the caps; or opted out, added to your internal do-not-call list and suppressed. Call activity and refusals write an exportable evidence trail, and completed call records are available in your workspace or through webhook, CSV, or API. 01 · The list 200,000 leads on your list Yours, captured or acquired, with provenance attested at import and cold outreach signed for under the waiver. You pick who gets called, and campaign attempts use your own Twilio subaccount and numbers. 02 · The gate Checked before any dial Numbers are normalized, compared with your internal do-not-call list, and kept inside the calling window and frequency caps. The gate fails closed, and refusals are recorded. 03 · The conversations Completed conversations become qualified outcomes Connected calls begin with an AI disclosure. The agent uses your rubric and product knowledge to capture the details your team needs. 04 · The outcomes Sorted, delivered, provable Qualified outcomes and dispositions are written to the relevant contact in your workspace, then made available to your stack through configured signed events, supported CSV exports, or API reads. data → check → call → qualify → outcome → webhook - ✓ NORMALIZE E.164 · malformed numbers rejected - ✓ INTERNAL_DNC your suppression list · re-checked at dial time - ✓ DIALABILITY the calling window and frequency caps - on your internal DNC → REFUSED · never dialed AI disclosure · before the first response synthetic transcript agent ▸ “Quick heads-up: I’m an AI assistant calling for Acme.” lead ▸ “Sure, what’s this about?” INTERESTED · OPTED IN with qualification and next-step details delivered to your stack call_evidence written · made or refused · exportable Proven, per call All of it without hiring a call team. Your pacing, your numbers, your rubric. Illustrative scenario. Your list, your rubric, your numbers. 01 · The list ### 200,000 leads on your list Yours, captured or acquired, with provenance attested at import and cold outreach signed for under the waiver. You pick who gets called, and campaign attempts use your own Twilio subaccount and numbers. 02 · The gate ### Checked before any dial Numbers are normalized, compared with your internal do-not-call list, and kept inside the calling window and frequency caps. The gate fails closed, and refusals are recorded. - NORMALIZE ✓ - INTERNAL_DNC ✓ - DIALABILITY ✓ - on your internal DNC → REFUSED · never dialed 03 · The conversations ### Completed conversations become qualified outcomes Connected calls begin with an AI disclosure. The agent uses your rubric and product knowledge to capture the details your team needs. agent ▸ “Quick heads-up: I’m an AI assistant calling for Acme.” synthetic transcript 04 · The outcomes ### Sorted, delivered, provable Qualified outcomes and dispositions are written to the relevant contact in your workspace, then made available to your stack through configured signed events, supported CSV exports, or API reads. - INTERESTED · OPTED IN → qualified outcome delivered to your stack - NOT NOW → documented, retry inside your caps - OPTED OUT → internal DNC, suppressed for good data → check → call → qualify → outcome → webhook call_evidence written · made or refused · exportable Proven, per call Illustrative scenario. Your list, your rubric, your numbers. Every check a number passes before it is dialed:compliance enforced in code. What the record it writes actually contains:the per-call evidence file. One platform · four jobs ## Put better calls to work across your business Build agents and workflows for revenue, operations, customer feedback, and inbound service. These workflows turn conversations into clear records and usable next steps. Revenue · Lead qualification ### Qualified pipeline from the leads you already have. Run a calling workflow that qualifies the leads you already have against your rubric. Interested leads can share the time they named, while disposition reasons give your reps clear context for their next step. - Qualified against your rubric - Time captured in their words - Reasons your reps can use A cold call, from first hello to next step Simulated demo · AI voices - agent ▸ “Hi Dale, being upfront: I’m an AI assistant calling for Crestway Software. We help trucking companies cut dispatch paperwork. Worth thirty seconds?” - lead ▸ “Uh, go ahead.” - agent ▸ “Who runs dispatch and invoicing today? Is that all in-house?” - lead ▸ “My office manager. Takes half her week, honestly.” - agent ▸ “That’s the exact fit. Here’s my offer: we build a short demonstration around your operation, your lanes, your paperwork, and send it over. Fair?” - lead ▸ “Yeah, I’d look at that.” - agent ▸ “Done. And once you’ve had a look, when’s good to walk through it - mornings any use to you?” - lead ▸ “Tuesday after nine works.” - agent ▸ “Tuesday after nine - noted, with everything else you told me. It’s in front of the Crestway team the second we hang up, and they’ll lock the time in. Thanks, Dale.” A professional smiles at her desk as morning campaign results come in (Illustrative footage.) Operations · Workforce calls ### Coordinate roster outreach from one workflow. Build confirmation calls around your sites, tone, and escalation language. Run roster outreach from one workspace and keep structured shift confirmations and decline reasons attached to the work. - Workflows for your operation - Roster outreach at scale - Shift confirmations and decline reasons A recruiter with a headset organizes tomorrow's roster (Illustrative footage.) Customer experience · Feedback calls ### Hear what customers say while the visit is fresh. Ask your questions in your voice after the transaction. Keep structured survey responses, call outcomes, and the context your team needs together for follow-up and reporting. - Your questions and follow-ups - Structured survey responses - CSAT and NPS views A customer at home answers a call (Illustrative footage.) Inbound · AI receptionist ### Give callers a useful first answer. Configure business hours, holiday rules, menus, routing, and AI answers on the number you already publish. Calls that need follow-up can create messages, tickets, callbacks, or updates to an eligible caller record. - Hours, holidays, and menus - Routes you control - Inbox follow-up A professional receives a customer call at their desk (Illustrative footage.) Published samples open with the AI disclosure callers hear. One business. Connected conversations. ## Your front desk, workforce and sales team. In one workspace. Think of a snow-service business preparing for a storm. It needs workers ready, customers supported and new applications followed up. Those are different conversations with a shared operating home. ### Get ready for the next shift. Call your roster to collect availability and shift responses. Keep tentative answers visible for your team to confirm. Keep track of worker responses. ### Give both a place to call. Configure reception for customer questions and worker support. Use your approved business knowledge and route requests to the right team. Keep track of messages and support requests. ### Find out whether you fit. Follow up with applicants using your qualification questions. Capture their needs and the next step they agree to. Keep track of qualification details. ### Start with one job. Add the next when you need it. Run multiple campaigns and add inbound reception to your outbound plan. Increase inbound connections and outbound sessions separately as more conversations need to happen at once. ### Bring the work from the systems you already use. Import a roster or lead list with CSV. Use the API and signed webhooks to connect your forms and send call outcomes back to your own tools. Illustrative operating example. Your team confirms staffing assignments and configures the calling permissions, business facts and integrations for each workflow. From your list to live calls ## Launch your first campaign in five guided steps Build and test in the free sandbox, then use the portal to choose a plan, complete production readiness, and connect the carrier account you already own. Ready workspaces activate production there; a carrier or compliance decision routes with the same context to review. - 01 Start from a no-tool draft Lead qualification, staffing outreach, and feedback calls: structured objectives, not a flow-builder maze. - 02 Make it yours Persona, objectives, your product knowledge, your voice. Drop in your pricing sheet, your FAQ, your product PDFs, and a knowledge pack turns them into what the agent actually says. - 03 Hear it Before you spend a cent, hear a call end to end as a simulated demo in AI voices, transcript published beside it, opening with the same disclosure the product plays on every call. - 04 Launch Choose a plan, accept the production order, add a payment method, and connect your own verified Twilio subaccount. The portal guides readiness and lets ready workspaces activate production; carrier or compliance decisions route with the same context to review. - 05 Records land Call records, webhooks, CSV, and your CRM, plus an evidence file for every call. A person checks their phone and smiles as campaign outcomes land (Illustrative footage.) Direct no-tool draft Speed-to-lead qualifier Structured objectives, not a flow-builder maze. Start with a direct draft and make the job yours. - objective: qualify against your rubric - objective: qualify and capture next-step details - guardrail: answer only from your knowledge Start from a direct draft and publish when it sounds like you. Completed conversations return structured outcome fields to your workspace and connected systems. Knowledge pack - your pricing sheet - your FAQ - your product PDFs - your objection notes persona Warm, direct, sounds like your best rep, in your voice,your vocabulary, your rules of engagement. Sandbox dial opening line, played before the first response, always “Hi, this is an AI assistant calling on behalf of your company. This call may be recorded…” Iterate on the voice until it sounds like your best rep, then launch. The AI disclosure is enforced in code, so it is structurally impossible to remove, on every plan. Pre-dial checks The three pre-dial checks the platform performs, run in order before any number is dialed: normalization, suppression against your own internal do-not-call list, and the calling window and frequency caps. Any check that cannot complete stops the call before it is placed, and the refusal is recorded with the value that caused it. - NORMALIZE E.164 · malformed rejected - INTERNAL_DNC your suppression list - DIALABILITY calling window · caps run report → reviewed & acknowledged list proceeds You read the run report before a single number is dialed, and every refusal in it is already an evidence row you can export. Live call → structured outcome - agent ▸ “What’s pushing you to look at this now?” - lead ▸ “We miss inbound leads on nights and weekends.” - agent ▸ “That’s exactly the job. When’s good for a walkthrough?” - lead ▸ “Thursday morning, say 10:30.” → intent and the time they named, captured as fields → available in the workspace, by configured event, CSV, or API → evidence file written · exportable Simulated example. Real outcomes are available in the workspace and through configured signed events, supported CSV jobs, or API reads. - 1 Start from a no-tool draft Lead qualification, staffing outreach, and feedback calls: structured objectives, not a flow-builder maze. Direct no-tool draft lead_qualification Speed-to-lead qualifier Structured objectives, not a flow-builder maze. Start with a direct draft and make the job yours. objective: qualify against your rubric objective: qualify and capture next-step details guardrail: answer only from your knowledge Start from a direct draft and publish when it sounds like you. Completed conversations return structured outcome fields to your workspace and connected systems. - 2 Make it yours Persona, objectives, your product knowledge, your voice. Drop in your pricing sheet, your FAQ, your product PDFs, and a knowledge pack turns them into what the agent actually says. Knowledge pack compiled your pricing sheet your FAQ your product PDFs your objection notes persona Warm, direct, sounds like your best rep, in your voice,your vocabulary, your rules of engagement. - 3 Hear it Before you spend a cent, hear a call end to end as a simulated demo in AI voices, transcript published beside it, opening with the same disclosure the product plays on every call. Sandbox dial the disclosure goes first opening line, played before the first response, always “Hi, this is an AI assistant calling on behalf of your company. This call may be recorded…” Iterate on the voice until it sounds like your best rep, then launch. The AI disclosure is enforced in code, so it is structurally impossible to remove, on every plan. - 4 Launch Choose a plan, accept the production order, add a payment method, and connect your own verified Twilio subaccount. The portal guides readiness and lets ready workspaces activate production; carrier or compliance decisions route with the same context to review. Pre-dial checks fails closed The three pre-dial checks the platform performs, run in order before any number is dialed: normalization, suppression against your own internal do-not-call list, and the calling window and frequency caps. Any check that cannot complete stops the call before it is placed, and the refusal is recorded with the value that caused it. NORMALIZE E.164 · malformed rejected INTERNAL_DNC your suppression list DIALABILITY calling window · caps +1 ••• ••• 4821 · on your internal DNC REFUSED · evidence recorded run report → reviewed & acknowledged list proceeds You read the run report before a single number is dialed, and every refusal in it is already an evidence row you can export. - 5 Records land Call records, webhooks, CSV, and your CRM, plus an evidence file for every call. Live call → structured outcome synthetic transcript agent ▸ “What’s pushing you to look at this now?” lead ▸ “We miss inbound leads on nights and weekends.” agent ▸ “That’s exactly the job. When’s good for a walkthrough?” lead ▸ “Thursday morning, say 10:30.” qualified · time captured · delivered M T W T F → intent and the time they named, captured as fields → available in the workspace, by configured event, CSV, or API → evidence file written · exportable Simulated example. Real outcomes are available in the workspace and through configured signed events, supported CSV jobs, or API reads. Step four in full, from verifying the subaccount to pointing a number and the signed callback:your carrier, your numbers. What lands in step five:the per-call evidence file. Your carrier, your rates ## Your Twilio rate is the rate you pay Every production call leaves your own Twilio account: your number, your rates, your relationship with the carrier. Twilio bills those minutes straight to you atthe price you already negotiated, with nothing of ours on top. It is also why our minute is priced the way it is: it buys the AI session, not the phone line. What connecting involves, from verifying the subaccount to pointing a number and the signed callback:your carrier, your numbers. The two ways to buy a phone call, compared as models rather than as vendors:BYO carrier vs resold minutes. ### Your account, your numbers Production calls leave from a Twilio subaccount you own: your numbers, your registrations, the rates you negotiated, and the caller-ID reputation those numbers earn. You connect it and verify it once in the portal. Vocapable runs on Twilio Media Streams, our own stack, and the account underneath stays yours to take anywhere. ### Your rate is the rate Twilio bills you for the minutes, directly, at the price you already negotiated. Our per-minute rate covers the AI session: the model, the speech, the time the agent spends talking. Two bills, both itemized, and the carrier one is between you and your carrier. ### Our numbers do one thing The platform’s own numbers carry one kind of traffic and one only: a paid test call to a number you proved you control. Every other destination is refused at the dialability gate, and the refusal lands in that attempt’s evidence. Code, not a policy page. You choose the list and the numbers it goes out on. We enforce what is set: the calling window, the frequency caps, the consent standard and the recording-consent class, applied on every attempt and recorded with the exact values that bound the call. That record is what you hand over when someone asks. Inbound calls, configured for your business ## Give callers a useful first answer. Put an AI receptionist on the Twilio number you already publish. Define your open and closed hours, holidays, timezone, menus, routes, and agent handoff in one answering flow that reflects how your business works. ### Your call flow, under your control Guide callers through menus, ringing your team, AI conversations, voicemail, spoken messages, or forwarding. A reserved inbound connection is used for the AI conversation, while the rest of the flow follows the steps you set. ### Follow-up your team can work When a caller needs help, the agent can capture a message, open a ticket, schedule a callback, or update a known caller's lead. The shared Inbox keeps those requests close to the call that created them. ### Predictable inbound capacity $250/mo includes 1 concurrent inbound connection, with $150/mo for each additional one. Answered calls add no per-call or per-minute Vocapable fee. Add Receptionist on its own or alongside an outbound plan. Your inbound connections and outbound sessions each keep their own capacity. Under every campaign ## A CRM built for one job: calling Campaigns need somewhere to put what they learned, and a general-purpose CRM with a phone bolted to the side is not it. Contacts carry your tags and your own fields. Lists are lists. Import maps your columns and hands back every row it refused. And a contact’s page is that person’s whole history across every campaign you have run, sitting beside the two facts that decide whether they get dialed again: what your do-not-call list holds, and what the dialability check returned. contact con_01J9…F2K Product UI · synthetic data Priya Natarajan Brightside Clinics · Ops Director - webinar-q3 - multi-site - sites: 4 purpose: marketing · America/Chicago, 8:12 PM · calling window and caps applied read live from your do-not-call list, not from a flag on the row captured Mar 4 · active Activity · all campaigns - Outbound call · qualified Q3 multi-site push · Tue 4:31 PM - Note · R. Cole “Wants pricing for two more sites before she commits.” - Note · after the call she asked for Thu 9:00 AM · written on the contact - Outbound call · voicemail Spring reactivation · Feb 9 Exports to CSV · every field here is a /v1 read ### What that page holds One contact, everything that has ever happened to them, and the two facts that decide the next dial. - Your tags, and the custom fields you invented, on the record itself - Every call across every campaign, with its disposition, the transcript turn by turn, and the recording your recording-consent class kept - Notes your reps typed, sitting beside the calls they came from - Lead status, and the rep the contact is assigned to, on the record rather than in a spreadsheet - What your internal do-not-call list holds for that number, when it landed there, and the call it came from - The dialability verdict for the purpose you pick, the reason it refused, and the policy values that decided - Consent records: class, source, when it was captured, when it expires ### Your columns, not ours Point an import at a CSV and map its columns onto the fields you keep. What survives lands in a list; what does not comes back row by row as JSON or as a CSV you can hand to whoever produced the file: the row number, the column, the value, and the reason. The consent attestation the import was submitted under, named signer and title, stays attached to the list it built. ### Filter like you mean it Every filter is a server filter and every page is a cursor, so a filtered view is the whole filtered set rather than the fifty rows your browser happened to be holding. Then act on what you filtered, up to a thousand contacts per request: add them to a list, take them off one, set lead status, or put the numbers on your internal do-not-call list. ### Take it all with you Contacts, attempts, calls, outcomes, your do-not-call list, your pre-dial results and your usage: nine resources, every one of them a CSV export you start yourself. Recordings come back as short-lived signed URLs, transcripts as ordered turns, dispositions with the whole history of how they changed. Leaving with your data is an afternoon, not a support ticket. Already have a system of record? Keep it. Drive the campaigns from your stack and let outcomes land where your team already works. Every screen above is built on the same /v1 your engineers get, so anything you have just seen, your code can do too. Metrics reported per campaign: Contact rate; Qualification rate; Demos requested; Disposition mix; Refusals, by reason; Cost per qualified outcome; Shift confirmations and decline reasons; Cost per completed call; Completed conversations; Structured survey responses; CSAT and NPS views. From calls to decisions ## Measure the work your calls create Review the outcomes, disposition trends, and usage behind your campaigns. The numbers come from the call, campaign, and usage records that produced them, so your team can trace a dashboard read back to the work. See the full reporting surface on the metrics we report. ### Lead programs - Contact rate - Qualification rate - Demos requested - Disposition mix - Refusals, by reason - Cost per qualified outcome ### Workforce calls - Contact rate - Shift confirmations and decline reasons - Cost per completed call ### Customer feedback - Contact rate - Completed conversations - Structured survey responses - CSAT and NPS views - Cost per completed call From conversations to next steps ## Calls become work your team can see Calls write their outcomes to records. Use the native Inbox for operational follow-up, or send the same context to the systems your team already uses through signed webhooks, exports, and the API. ### A call record you can inspect Open a call with its outcome and disposition history, plus available evidence such as a transcript, recording, or voicemail. The context stays with the conversation instead of getting lost in a handoff. ### Follow-up in one place Messages, tickets, and callback requests give people a clear next action. When a caller matches an eligible record, the receptionist can update it too; assign and resolve requests from the operational Inbox. ### Structured outcomes Keep lead qualification, structured shift confirmations and decline reasons, and survey responses attached to the work. Campaign and survey views make progress, disposition trends, CSAT, and NPS easier to review. ### Connect it to your systems Receive signed call and outcome events, inspect or retry delivery, and use the /v1 API or CSV exports when your team needs the data in its own workflow. Records that travel with the work ## Calls your team can explain and act on A call record captures the checks that ran, the signature behind the list, and what happened on the line. Use it to continue the work, investigate a problem, or export the evidence for someone who needs it. attempt att_01J9…K7Q Synthetic data Every line is written by the platform as the call happens, with the signature that admitted the list attached to it. The file exports exactly as you see it. Exportable evidence attempt att_01J9…M2X Synthetic data Your suppression list is checked at import and again at dial time, so the policy decision remains attached to the record. Refused, and proven ### Policy controls that run with the call Configure the calling policy your team needs, including normalization, your own do-not-call list, calling windows, and frequency caps. The AI disclosure and opt-out handling are recorded as the call runs. ### A record your team can use Made or refused, calls produce an exportable record: the checks that ran, the signature that admitted the list, and what happened on the line. It is written as the call happens, ready for the next person who needs it. ### Clear carrier ownership A published rate card and one bill. Every plan includes the compliance engine at $0, while the carrier leg stays on your own Twilio account at the rates you negotiated. How each control works, one by one:the compliance engine. What one call's record actually holds, with a sample file to open:the per-call evidence file. Every control on this page in one place:the Trust Center. Your CRM, or ours ## Run it from the dashboards, or never open them. ### Workspace Your contacts, imported or sent through the API, and then everything from the dashboards: lists, campaigns, live boards, and a CRM built for one job, where every call that ever touched a record sits on one timeline with its disposition, transcript, notes and human follow-up context. Vocapable is your calling system of record. ### Headless Keep your stack as the system of record. Product resources use the/v1 API, so your code can create contacts, place calls on your own Twilio, and receive enabled events at your configured signed webhook. Signup, onboarding, and first-key setup run in the browser. Same engine, same compliance floor, same pricing. Mix the two freely, per team, per campaign. Read the docs → ## Published prices and usage rates. Start free in the sandbox. Three outbound workload classes, starting at $349/mo with AI usage from $0.13/min - enterprise plans as low as $0.07/min.Inbound answering is priced separately, from $250/mo. Every rate we bill is printed on the pricing page, so the first invoice holds no surprises. Your Twilio minutes stay on your own account at your own rate. The sandbox is free, you pay when you go live, and the compliance engine is $0 on every plan. Or put your own numbers in andprice your own campaign against the published rates, before you talk to anyone. ## Your list is sitting there. Call it. The sandbox is free: your real use case, your real material, the whole dial path running with every gate live. And you hear the simulated demo calls before you sign anything. --- # Legal | Vocapable > The Vocapable legal document set: MSA, Telecom and Test Calls Addenda, AUP, AI Disclosure, DPA, Cold Outreach Waiver, Privacy Policy, and Website Terms. URL: https://www.vocapable.com/legal # Legal Read exactly what you are agreeing to, before you agree to it. Every document below is published from the same source text the product serves for acceptance, so the version on this page is the version you sign, hash for hash, and the two cannot drift apart. ## The platform agreement The Master Services Agreement and every document it pulls in. You accept this set in-product before your first call, and you can read all of it here first. ### Master Services Agreement The core agreement governing use of the Vocapable platform. Incorporates the addenda, AUP, AI disclosure, DPA, and privacy policy by reference. ### Telecom Compliance Addendum Your calling-law obligations for consent, scrubbing, disclosures, and revocation. Controls over the MSA body where they conflict. ### Platform Test Calls Addendum The only permitted use of platform-owned numbers: verified test calls to numbers you control. ### Acceptable Use Policy What may never run on the Service, and what happens when it does. ### AI Technology Disclosure A plain-language notice of what our AI is, what can go wrong, and what relying on it means. ### Data Processing Addendum CCPA/CPRA service-provider terms for the personal information we process on your behalf, plus the two retention carve-outs that survive a deletion request. ## Signed when you elect it You sign this one only if you choose to run the outreach it covers, and it controls over the other documents for that outreach. Nobody is bound by it at signup. ### Cold Outreach & Data Provenance Waiver The document a customer signs to run cold outreach, calls to purchased, rented, or appended lists, and artificial-voice voicemail: their own list, their own carrier, and their own responsibility for the lawfulness of every call. It controls over the other documents for that outreach. ## Site policies How this website behaves and how we handle personal information. ### Privacy Policy How we handle personal information: yours as a customer, and the data we process on customers’ behalf. ### Website Terms of Use Terms for this website only. Use of the platform itself is governed by the Master Services Agreement. Want the controls behind these documents? The Trust Center has the four-control floor, the security posture, and who holds what. Questions about any of these documents: support@capstralabs.com. --- # AI Technology Disclosure | Vocapable > Vocapable AI Technology Disclosure: a plain-language notice of what our AI voice agents are, what can go wrong, and what relying on them means. URL: https://www.vocapable.com/legal/ai-disclosure ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # AI Technology Disclosure Version: v1.0 · Referenced by the Vocapable Master Services Agreement §4. This notice explains, in plain language, what Vocapable’s AI technology is and what it means to rely on it. It is part of your agreement with us. ## What the technology is Vocapable’s voice agents are powered by artificial intelligence: large language models that generate what the agent says, speech-recognition models that transcribe what people say, and speech-synthesis models that produce the agent’s voice. Some of these models are built and operated by us; others are provided by third parties. These systems are probabilistic. They do not follow a fixed script the way a recording does — they generate responses in the moment. That is what makes them useful, and it is also why they can be wrong. ## What can go wrong You should expect, and plan for, all of the following — before, during, and after calls: - An agent may say something inaccurate, incomplete, or invented (“hallucination”), even about your own product, prices, or policies, and even when your knowledge sources are correct. - An agent may misunderstand the person it’s speaking with, mishear words, or respond to something that wasn’t said. - An agent may deviate from your configuration — your prompts, objectives, and guardrails reduce this but cannot eliminate it. - Transcripts, summaries, extracted data, and scores may contain errors. - Behavior may change over time as models are updated, including third-party models we do not control. - Calls may fail, drop, or degrade for technical or carrier reasons. - The same input can produce different outputs on different runs. This is an experimental, rapidly evolving technology under ongoing development. We test extensively — every agent version passes an evaluation gate before it can be published — but testing reduces error rates; it does not make errors impossible. ## What we ask of you (and what you agreed to) - Review outputs before relying on them. Do not treat anything an agent said, or any transcript or summary, as verified fact without checking it. - Decide where human review belongs in your workflow. You know your use case; we don’t. - Never use the Service as the sole basis for decisions with legal, financial, medical, employment, or similarly significant effects on a person. - Tell people they’re talking to an AI. Every call opens with a non-removable AI disclosure — that’s a floor we enforce, not a ceiling. Your own use case may require more under applicable law. - Configure carefully. Your prompts, knowledge packs, and guardrails are yours; the agent speaks on your behalf and at your direction. ## What this means legally As set out in the Master Services Agreement: the Service and its outputs are provided “as is”; we do not warrant their accuracy; you are responsible for evaluating their fitness for your use; we are not responsible for third-party models and providers; and, as between you and us, you bear the risk of adverse outcomes caused by output errors, subject to the MSA’s limitation-of-liability and indemnification provisions. If any part of this notice conflicts with the MSA, the MSA controls. --- # Acceptable Use Policy | Vocapable > Vocapable Acceptable Use Policy: what may never run on the Service, including scraped contact data, deception, harassment, and network abuse, and the waiver that cold outreach and acquired contact lists require. URL: https://www.vocapable.com/legal/aup ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Acceptable Use Policy Version: v1.1 · Incorporated into the Vocapable Master Services Agreement. (v1.1, 2026-08-15: Section 1, item 2 rewritten — cold outreach and acquired contact lists are permitted only under an executed Cold Outreach & Data Provenance Waiver, and scraped or harvested data remains prohibited in all cases with no waiver available.) This Acceptable Use Policy (“AUP”) applies to everyone who uses the Service. Customer is responsible for its Users’ and downstream clients’ compliance with this AUP. ## 1. Prohibited Uses You may not use the Service to: - Break the law. Violate any law, regulation, governmental order, industry standard, or telecommunications carrier requirement in any applicable jurisdiction — including all Telecom Laws (see the Telecom Compliance Addendum). - Cold-call outside an executed waiver. Place outbound marketing or sales calls to any person for whom you do not hold documented, unrevoked consent of the legally required class, or call contacts obtained from a list seller, broker, or data appender — except where you have executed the Cold Outreach & Data Provenance Waiver and the calls are Elected Outreach under it. That Waiver permits the conduct on the Service and allocates its risk to you. It is not a determination that any call is lawful, Vocapable makes no such determination, and it moves no consent, registry, screening, disclosure, or record-keeping obligation from you to Vocapable. Every other provision of this AUP continues to apply to Elected Outreach in full. Contact data obtained by scraping, harvesting, or any other unauthorized extraction is prohibited on the Service in all cases, however it would reach the Service, and no waiver of that prohibition is available. - Deceive. Impersonate any person or entity; misrepresent who is calling or why; configure an agent to deny being an AI, to claim to be human, or to defeat or omit the AI disclosure; falsify caller ID or message origin; or engage in fraud, phishing, pretexting, or social engineering. - Harm or harass. Threaten, harass, abuse, defame, stalk, or intimidate any person; place calls intended to annoy; or contact any person who has opted out. - Target the vulnerable or the regulated. Conduct campaigns directed at minors; collect payment-card data (PCI), protected health information (HIPAA), or conduct debt collection (FDCPA/Reg F) absent a separately executed plan expressly permitting it; place political campaign robocalls; or call emergency services. - Abuse the network. Generate calling patterns that harm number reputation or the telephone network, including high volumes of unanswered attempts from a number, abnormally short average call durations, sequential/enumerated dialing, or traffic that generates an unacceptable level of complaints, spam labeling, carrier blocking, or traceback requests. - Abuse the platform. Circumvent or probe any technical control (including the compliance floor, the platform test-call restriction, rate limits, or spend caps); share or resell access except as permitted in writing; reverse engineer the Service; use the Service to build or train a competing model or dataset; introduce malware; or interfere with other customers. - Misuse voices. Clone or synthesize a voice without the verified, documented consent of its owner, or use a synthetic voice to impersonate a real person without authority. ## 2. Monitoring and Enforcement Vocapable may monitor usage signals (call patterns, dispositions, complaint and traceback rates, content flags from automated review) to operate the platform, meet its own carrier and legal obligations, and enforce this AUP. Vocapable may audit, meter, throttle, block, suspend, or terminate any campaign, agent, number, or account it reasonably believes violates this AUP or creates risk to called parties, the network, or the platform — without prior notice where necessary. Violations may be reported to carriers or authorities where required. Enforcement actions and their basis are logged. ## 3. Reporting Report suspected abuse of the Service to support@capstralabs.com with the subject “Abuse report.” We investigate every report. --- # Cold Outreach & Data Provenance Waiver | Vocapable > The Vocapable Cold Outreach & Data Provenance Waiver: what a customer signs to run cold outreach, to call purchased, rented, or appended lists, or to leave artificial-voice voicemail, and where it puts responsibility for those calls. URL: https://www.vocapable.com/legal/cold-outreach-waiver ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Cold Outreach & Data Provenance Waiver Version: v1.0 · Incorporated into the Vocapable Master Services Agreement upon execution. This Cold Outreach & Data Provenance Waiver (this “Waiver”) is entered into betweenCapstra Labs, LLC, doing business as Vocapable (“Vocapable”), and the entity accepting this Waiver (“Customer”). It supplements the Vocapable Master Services Agreement (the “MSA”), the Acceptable Use Policy, and the Telecom Compliance Addendum, and applies where Customer elects to conduct any of the following (each, “Elected Outreach”): (a) calls to persons from whom Customer has not obtained prior consent to be called (“Cold Outreach”); (b) calls to contacts obtained from a third-party list seller, broker, or data appender (“Acquired Data”); or (c) the delivery of artificial-voice voicemail messages. If this Waiver conflicts with any other document, this Waiver controls with respect to Elected Outreach. The individual accepting represents that they have authority to bind Customer. ## 1. Customer’s Program, Customer’s Data, Customer’s Carrier 1.1 Elected Outreach is Customer’s program. Customer — not Vocapable — selects the contacts, supplies the data, defines the campaign, and originates every call on Customer’s own carrier account and phone numbers. Vocapable provides software. Vocapable does not sell, rent, source, or supply contact data, does not select whom Customer calls, and does not direct the timing, volume, or content of Customer’s campaigns. 1.2 Data provenance. Customer represents and warrants that all Acquired Data was obtained lawfully, under a valid agreement with the data supplier that permits its use for Customer’s calling programs, and that no contact data supplied to the Service was obtained by scraping, harvesting, or any unauthorized extraction. Scraped or unauthorized-extraction data remains prohibited on the platform in all cases. 1.3 Carrier relationship. All production Elected Outreach rides Customer’s own carrier account. Customer is solely responsible for its carrier relationship, including number registrations, A2P and STIR/SHAKEN attestations, and compliance with its carrier’s acceptable-use policies. Suspension or termination by Customer’s carrier is not a Service failure and gives rise to no refund, credit, or claim against Vocapable. ## 2. Sole Responsibility for Lawfulness 2.1 Customer is solely responsible for the lawfulness of every call and voicemail placed through the Service as Elected Outreach, including without limitation compliance with: the Telephone Consumer Protection Act (TCPA) and its artificial-voice and prior-express-written-consent rules; the FTC Telemarketing Sales Rule; federal, state, and internal Do-Not-Call registries; state telemarketing and “mini-TCPA” statutes; call-time, frequency, identification, and disclosure requirements; and any consent, notice, or record-keeping obligation applicable to Customer’s contacts, jurisdictions, or industry. 2.2 Platform controls are not legal clearance. The Service applies certain floor controls to every call — including Do-Not-Call scrubbing against lists the platform can access, jurisdiction rules for covered states, an AI-disclosure opener, and opt-out handling. Customer acknowledges that these controls are operational safeguards only: they do not verify Customer’s consent basis, do not cover every list or jurisdiction, do not constitute legal advice, and do not make any call lawful. Vocapable makes no representation or warranty that any Elected Outreach is lawful, and Customer has not relied on any such representation. 2.3 No legal advice. Nothing in the Service, its documentation, its screening questions, or its marketing is legal advice. Customer confirms it has had the opportunity to consult its own counsel regarding Elected Outreach. ## 3. Assumption of Risk; Waiver of Claims 3.1 To the maximum extent permitted by applicable law, Customer knowingly and voluntarily assumes all risk arising from Elected Outreach, including regulatory penalties, private claims (including class actions), carrier enforcement, and number or brand reputation damage, whether or not foreseeable. 3.2 To the maximum extent permitted by applicable law, Customer waives any right to assert, claim, demand, or recover from Vocapable any loss, fee, refund, credit, or damages — on any legal or equitable theory, including negligence, breach of warranty, failure of consideration, frustration of purpose, impossibility, impracticability, or unjust enrichment — arising from or relating to: (a) Elected Outreach and its consequences; (b) any change in law, regulation, or carrier policy that restricts, burdens, or prohibits Elected Outreach; (c) Vocapable’s suspension, restriction, or termination of Elected Outreach under Section 4; or (d) any third-party or regulator claim concerning Customer’s campaigns or data. ## 4. Vocapable’s Right to Refuse and Suspend 4.1 Vocapable may, at any time and without liability, refuse, restrict, suspend, or terminate any Elected Outreach, campaign, or account feature where Vocapable reasonably believes it is unlawful, violates the AUP, endangers platform integrity or carrier standing, or is the subject of complaints, regulator inquiry, or carrier action. Fees already paid or accrued are not refunded or credited on account of such action. ## 5. Indemnification 5.1 Customer will defend, indemnify, and hold harmless Vocapable, its affiliates, and their officers, directors, employees, and agents from and against any and all claims, actions, investigations, demands, losses, liabilities, damages, settlements, penalties, fines, costs, and expenses (including reasonable attorneys’ fees) arising from or relating to: (a) Elected Outreach; (b) Customer’s contact data, its provenance, or its use; (c) Customer’s violation of any law, regulation, or carrier policy; or (d) Customer’s breach of this Waiver. Vocapable may participate in the defense with counsel of its choosing at Customer’s expense, and Customer will not settle any claim imposing obligations on Vocapable without Vocapable’s prior written consent. ## 6. General 6.1 This Waiver survives the termination or expiration of the MSA with respect to Elected Outreach conducted while it was in effect. 6.2 If any provision of this Waiver is held unenforceable, it will be enforced to the maximum extent permissible, and the remainder will continue in full force. A holding that limits one clause of Section 3 or 5 does not disturb the others. 6.3 Customer’s acceptance of this Waiver is recorded with the accepted version, a cryptographic hash of the exact text presented, the signer’s name, title, and email, and a server-recorded timestamp and network address, and constitutes Customer’s execution of this Waiver. --- # Data Processing Addendum | Vocapable > Vocapable Data Processing Addendum: the CCPA/CPRA service-provider terms governing personal information we process on a customer's behalf, and the two retention carve-outs. URL: https://www.vocapable.com/legal/dpa ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Data Processing Addendum Version: v1.0 · Incorporated into the Vocapable Master Services Agreement. ## 1. Roles and Scope 1.1 For personal information in Customer Content and call data (contact records, call audio, transcripts, consent records, outcomes), Customer is the business/controller and Vocapable is the service provider/processor, processing solely to provide the Service per the MSA and Customer’s documented instructions (the configuration and use of the Service being such instructions). 1.2 For Vocapable’s own account, billing, security, and usage telemetry data, Vocapable is an independent business/controller per its Privacy Policy. ## 2. Vocapable’s Obligations Vocapable will, with respect to personal information processed on Customer’s behalf (the CCPA-required certifications of Cal. Civ. Code § 1798.100(d) and 11 CCR § 7051): - process it only for the limited and specified purpose of providing the Service described in the MSA, and not for any other purpose; - not sell it or share it (as those terms are defined by the CCPA/CPRA); - not retain, use, or disclose it outside the direct business relationship between the parties or for any purpose (including a commercial purpose) other than providing the Service, except as permitted by the CCPA and its regulations; - not combine it with personal information received from other sources, except as permitted by the CCPA and its regulations; - comply with all applicable obligations under the CCPA and provide the same level of privacy protection required of businesses; - notify Customer if it determines it can no longer meet its obligations under the CCPA, and grant Customer the right, upon such notice, to take reasonable and appropriate steps to stop and remediate unauthorized use of personal information; - permit Customer to take reasonable and appropriate steps to help ensure Vocapable uses personal information consistently with Customer’s CCPA obligations, satisfied by the audit provision in Section 5; - impose confidentiality obligations on personnel with access; - maintain reasonable technical and organizational security measures (tenant isolation via row-level security, encryption in transit and at rest, credential vaulting, access logging); - provide breach notice to Customer without undue delay after confirmation; and - provide reasonable assistance with data-subject requests and, at Customer’s expense, with assessments. ## 3. Subprocessors Customer generally authorizes subprocessors used to provide the Service, currently:Amazon Web Services (hosting and storage), Twilio (platform test-call telephony, and messaging where configured), Stripe (payments), xAI and other model providers as configured (language-model inference), and Resend (transactional email). Vocapable provides the current list on request to support@capstralabs.com, will provide notice of additions, and remains responsible for subprocessors’ performance. Where Customer brings its own provider credentials (BYO Twilio, BYO model keys), that provider processes under Customer’s own agreement with it and is not a Vocapable subprocessor. ## 4. Deletion and Retention — the Compliance Carve-Outs 4.1 On termination or verified deletion request, Vocapable deletes or de-identifies personal information within 60 days, except: - Compliance evidence — consent records, scrub results and verdicts, disclosure and opt-out evidence, attestations, and per-call policy snapshots — is retained for at least five (5) years notwithstanding any deletion request, because it is the legal-defense record of both parties and is retained as a legal obligation and for the establishment and defense of legal claims. - Suppression data — the do-not-call ledger — is retained indefinitely, because deleting it would cause the very contact the record exists to prevent. Suppression entries are minimized to what suppression requires. 4.2 Deleted contacts are tombstoned (no new calls may start to them) with the carve-outs above preserved. ## 5. Audit Vocapable will make available information reasonably necessary to demonstrate compliance with this DPA, including summaries of third-party audits if and when such audits exist (none is represented to exist today), no more than annually absent a confirmed incident. ## 6. Order of Precedence This DPA controls over the MSA for personal-information processing. Processing takes place in the United States. The Service is offered to U.S. businesses for U.S. calling; if the parties later agree to processing that requires additional transfer mechanisms or state-law addenda, they will execute them before that processing begins. --- # Privacy Policy | Vocapable > Vocapable Privacy Policy: how we handle personal information, covering customer account data and the call data we process on customers' behalf. URL: https://www.vocapable.com/legal/privacy ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Vocapable Privacy Policy Version: v1.0 · Effective: August 2, 2026 Capstra Labs, LLC, a California limited liability company, d/b/a Vocapable (“Vocapable,” “we”) provides a business platform on which our customers operate AI voice agents. This policy explains how we handle personal information. It covers our website and the accounts of our customers. For the people our customers call: our customer decides whom to call and why; we process call data on that customer’s behalf as a service provider. Requests about a specific call are best directed to the business that called you — we will route any request we receive to the responsible customer, and we honor do-not-call requests immediately and permanently on our platform. ## What we collect In the last 12 months we have collected, or will collect once the Service launches, the following categories of personal information (California Civil Code § 1798.140): - Identifiers and professional information (from customers and prospective customers): name, work email, job title, company, business phone, website, and the answers given during signup, including the compliance questionnaire, which we retain as a signed attestation. - Commercial information: subscription, order, and payment records. Payment cards are handled by Stripe; we never store card numbers. - Internet activity: server logs, usage metering, approximate location from IP address, and the minimal analytics described below. - Audio and derived records (on customers’ behalf, as a service provider): contact records our customers import, call audio and transcripts where recording is enabled and lawful, call outcomes, and the consent and compliance evidence connected to calls. - Website data: demo requests and contact forms (forwarded, not stored on the website itself). ## How we use it To provide, secure, meter, and bill the Service; to enforce our Acceptable Use Policy and meet carrier and legal obligations (including do-not-call suppression and call evidence retention); to support customers; to send transactional and account email; and to improve the Service. We do not use customer content or call recordings to train foundation models. We do not sell or share personal information as those terms are defined by the CCPA/CPRA, we have not done so in the preceding 12 months, and we do not use advertising trackers on the platform. We do not knowingly collect information from minors under 16 and have no actual knowledge of selling or sharing their information. ## Retention Account data for the life of the account plus 2 years; service data per customer configuration and the Data Processing Addendum — including its compliance carve-outs: consent, scrub, disclosure, and opt-out evidence is retained at least five years, and do-not-call suppression records are retained indefinitely, because deleting a do-not-call record would cause the contact it exists to prevent. Signup applications that never become accounts are retained up to 2 years for follow-up and legal-defense purposes, then deleted. ## Sharing Service providers listed in our DPA subprocessor list (hosting, telephony, payments, model inference, transactional email); authorities where required by law; a successor in a corporate transaction. Never data brokers. ## Your California privacy rights If you are a California resident, you may request that we: disclose the categories and specific pieces of personal information we have collected about you, the categories of sources, our purposes, and the categories of third parties to whom we disclose it;delete personal information we collected from you (subject to the legal-obligation and legal-claims exceptions above, which cover our compliance evidence and do-not-call records); and correct inaccurate personal information. We do not sell or share personal information, so there is no opt-out to exercise, and we do not use or disclose sensitive personal information for purposes requiring a right to limit. To exercise these rights, email support@capstralabs.com with the subject “Privacy request.” We verify requests by confirming control of the email address associated with the information and, where needed, asking for information that matches our records. You may use an authorized agent with signed permission. We respond within 45 days (extendable once by 45 days with notice) and will not discriminate against you for exercising any right. If we deny a request, you may appeal by replying to our response; we will answer the appeal within 45 days. Where the information belongs to one of our customers’ call recipients, the CCPA directs the request to the business that decided to collect it; we forward such requests to the responsible customer and assist them as their service provider. ## Do Not Track Our website does not respond to browser Do Not Track signals; we do not track visitors across third-party websites. Because we do not sell or share personal information, we do not process opt-out preference signals for those purposes. ## Analytics Our website uses privacy-respecting analytics without cross-site tracking, and shows a consent banner where required. ## Changes; contact We will post any changes here with a new version and effective date. Questions and requests: support@capstralabs.com. --- # Telecom Compliance Addendum | Vocapable > Vocapable Telecom Compliance Addendum: the customer's calling-law obligations for consent, scrubbing, disclosures, and revocation. URL: https://www.vocapable.com/legal/telecom-addendum ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Telecom Compliance Addendum Version: v1.2 · Incorporated into the Vocapable Master Services Agreement. (v1.2, 2026-08-15: Sections 2.3, 2.4 and 9.3 — acquired contact lists and outreach without a consent basis are permitted only under an executed Cold Outreach & Data Provenance Waiver, which carries its own records obligation; scraped or harvested data remains prohibited in all cases. v1.1, 2026-08-11: Section 3.3 added — the Service performs no registry, litigator, reassigned-number, or line-type screening, and says so rather than implying otherwise.) This Addendum states Customer’s obligations for every telephone call and message placed or sent through the Service. Capitalized terms have the MSA’s meanings. “Telecom Laws” means all laws, regulations, orders, and carrier and industry requirements applicable to Customer’s calls and messages, including the Telephone Consumer Protection Act (47 U.S.C. § 227) and FCC rules and declaratory rulings thereunder (including FCC 24-17, under which AI-generated voices are “artificial” voices), the Telemarketing Sales Rule, the Telemarketing and Consumer Fraud and Abuse Prevention Act, the Truth in Caller ID Act, and all state telemarketing, automated-calling, artificial-voice, AI-disclosure, call-recording, do-not-call, and privacy laws. ## 1. Allocation of Responsibility 1.1 Customer is solely responsible for compliance with all Telecom Laws applicable to its specific use case. Customer, and not Vocapable, determines who is called, when they are called, and what its agents are configured to say, and Customer is the party on whose behalf all calls are initiated. 1.2 The Service’s compliance features (scrubbing, gating, windows, disclosure, opt-out handling, evidence) support Customer’s compliance program but do not constitute legal advice, do not guarantee any call is lawful, and do not shift any obligation to Vocapable. A feature’s existence is not a representation that it satisfies any law. ## 2. Consent 2.1 Per-call continuing representation. Except as Section 2.4 provides for Elected Outreach, Customer represents and warrants, on a continuing basis and as to each individual call and message, that before initiating it Customer has obtained and holds valid, documented, and unrevoked consent of the called party of the type required under applicable law for that communication, including: - Prior express written consent (“PEWC”) where the call constitutes telemarketing or advertising and uses an artificial or prerecorded voice (which includes every AI-voice call on this Service) or is made to a wireless number using regulated technology. PEWC must (a) be in a signed writing (electronic signatures under E-SIGN/UETA are permitted); (b) be clear and conspicuous; (c) disclose that the consumer agrees to receive calls made with an artificial or prerecorded voice from or on behalf of the specific seller; (d) not be obtained through a pre-checked box; and (e) state that consent is not a condition of any purchase. - Prior express consent for informational, non-marketing calls, as applicable. - An applicable established business relationship or transactional basis only where, and only to the extent, applicable law recognizes it for the call type in question. 2.2 Best practice (contractual, not a statement of law): Customer should obtain consent that names the specific seller on whose behalf calls will be made. (The FCC’s one-to-one consent rule was vacated in 2025; this clause is therefore stated as best practice rather than as a statement of current law.) 2.3 Acquired lists require an executed waiver; scraped data is prohibited outright.Customer may not import or call purchased, rented, appended, or otherwise third-party-sourced contact lists unless Customer has executed the Cold Outreach & Data Provenance Waiver, in which case those contacts are Acquired Data under that Waiver and Customer’s provenance representation in Waiver Section 1.2 governs them. Every other imported contact must have a consent provenance Customer itself captured and can evidence, attested at import. Contact data obtained by scraping, harvesting, or any other unauthorized extraction may never be imported or called, with or without the Waiver, and no waiver of that prohibition is available. 2.4 Elected Outreach. Where Customer has executed the Cold Outreach & Data Provenance Waiver, Section 2.1’s per-call consent representation does not apply to the Elected Outreach that Waiver covers — calls to persons from whom Customer holds no prior consent, calls to Acquired Data, and the delivery of artificial-voice voicemail — and is replaced, for those communications only, by Waiver Section 2.1: Customer is solely responsible for the lawfulness of each such call and message, including every consent, registry, timing, frequency, identification, disclosure, and record-keeping requirement applicable to it. Every other Section of this Addendum continues to apply to Elected Outreach in full, including Sections 3 through 12, and Section 9.3 states what Section 9 requires of Elected Outreach records. The Waiver permits Elected Outreach on the Service and allocates its risk to Customer; it is not a representation by Vocapable that any Elected Outreach call or voicemail is lawful, and Vocapable makes none. The Service’s compliance features apply to Elected Outreach exactly as they apply to every other call, and Section 1.2 governs what they are worth. ## 3. Do-Not-Call 3.1 Customer will scrub its calling lists against the National Do Not Call Registry at least every 31 days (pre-dial suppression strongly preferred), against applicablestate do-not-call registries, and against Customer’s own internal suppression list, before any call subject to DNC rules. 3.2 The Service maintains and enforces an internal do-not-call list for Customer’s account as a non-disableable platform control. That control supplements and does not replace Customer’s own DNC obligations. 3.3 The Service does not screen Customer’s numbers against the federal or any state do-not-call registry, against any litigator or serial-complainer list, against the FCC Reassigned Numbers Database, or against line-type or wireless-ported data. Vocapable subscribes to no compliance data bureau and performs none of these checks, on any plan, at any price, for any customer. Customer’s own upstream screening is the only screening these sources receive, and it must meet or exceed Section 3.1 on the cadence Customer attests. Every production account executes the Upstream Contact-Screening Attestation before production activation, and each unperformed check is recorded in the Service’s evidence trail as an election under that attestation, naming its signer — never as a check that passed. ## 4. Revocation and Opt-Out 4.1 A called party may revoke consent by any reasonable means. Customer will honor revocations within ten (10) business days of receipt and will maintain its internal suppression list accordingly. The Service’s in-call opt-out detection writes revocations to the internal do-not-call list immediately and this cannot be disabled; Customer must honor revocations received through any other channel (email, web, verbal to Customer’s staff) within the same period. 4.2 Every artificial-voice telemarketing call must provide an automated, interactive opt-out mechanism available during the call. ## 5. Calling Windows and Frequency 5.1 No calls before 8:00 a.m. or after 9:00 p.m. in the called party’s local time zone, and no calls outside any narrower window imposed by applicable state law. Customer is responsible for state-specific windows, holiday restrictions, and frequency caps applicable to its campaigns, and for the accuracy of any time-zone data it supplies. ## 6. Identification and Disclosure 6.1 Each outbound call must disclose, promptly at the beginning: the identity of the business on whose behalf the call is made; the purpose of the call; and a callback number or contact method. Calls made with AI-generated voices must include any AI or artificial-voice disclosure required by applicable law. The Service’s non-removable AI-disclosure opener supplements, and does not replace, these obligations. ## 7. Caller ID; Carrier Obligations 7.1 Customer will not transmit misleading or inaccurate caller ID (Truth in Caller ID Act), will only present numbers it has authority to use, and will not engage in “neighbor spoofing.” 7.2 For production traffic (Customer’s own carrier account), Customer is responsible for: its carrier registrations, including Robocall Mitigation Database registration where applicable; STIR/SHAKEN attestation; A2P/10DLC registration for messaging; number reputation and spam-label remediation; and compliance with its carrier’s policies, including Twilio’s Acceptable Use Policy and Voice Services Policy, which are incorporated into this Addendum by reference for traffic carried on Twilio. 7.3 Traceback cooperation. Customer will respond substantively to any industry traceback (e.g., ITG), carrier, or regulatory inquiry regarding its traffic withinforty-eight (48) hours of Vocapable’s notice, and will cooperate with Vocapable and third parties to determine the origin of any call. ## 8. Recording 8.1 If Customer enables call recording, Customer is responsible for disclosing the recording and obtaining all consents required by applicable law, including inall-party-consent jurisdictions such as California (Cal. Penal Code §§ 632, 632.7), where every party to the call must consent to recording. The Service’s recording-consent configurations are tools, not advice, and enabling recording is Customer’s election. ## 9. Records; Production on Demand 9.1 Customer is solely responsible for obtaining, documenting, and retaining evidence of valid consent before initiating calls, capturing at minimum: timestamp; capture channel; the exact disclosure language presented; the affirmative act taken; and the identifier (IP, device, signature) of the consenting party. Customer will retain these records for at least five (5) years or longer where required, subject to litigation hold. 9.2 Upon Vocapable’s written request, Customer will produce the consent evidence for any called number within five (5) business days. Vocapable may make such requests to satisfy its own carrier obligations (Twilio may demand proof of consent from Vocapable). Failure to produce is a material breach and grounds for immediate suspension of the affected campaigns. 9.3 Elected Outreach records. For a call made as Elected Outreach, the records Customer must retain under Section 9.1 and produce under Section 9.2 are the provenance and program records for the number called: where the contacts are Acquired Data, the supplier, broker, or appender agreement relied on, its permitted-use terms, and the acquisition date and source; how the number was otherwise obtained where it is not Acquired Data; any consent, opt-out, or suppression data received with the contacts; any consent evidence Customer does hold; the campaign and script under which the number was called; and Customer’s executed Waiver. Section 9.1’s five-year retention and Section 9.2’s five-business-day production obligation apply to those records without change, and failure to produce them is a material breach and grounds for immediate suspension of the affected campaigns. ## 10. Prohibited Sectors and Uses 10.1 Absent a separately executed plan expressly permitting it, Customer will not use the Service: for calls involving protected health information (HIPAA); for payment-card data capture (PCI DSS); for debt collection subject to the FDCPA or Regulation F; for political campaign robocalls; for emergency or life-safety communications; or for any call to emergency services numbers. ## 11. SMS 11.1 All messaging follows the consent, opt-out (STOP), identification, and content rules of applicable law and carrier policy, including Twilio’s Messaging Policy(incorporated by reference for Twilio-carried messages). Revocation on one channel is honored across channels as required by law; the Service applies revocations across voice and SMS as a platform control. ## 12. Remedies 12.1 Breach of this Addendum is a material breach of the MSA. Vocapable may exercise its monitoring, suspension, and termination rights (MSA §12) immediately and without prior notice for suspected violations, and Customer’s indemnity (MSA §10.1) expressly covers this Addendum. --- # Master Services Agreement | Vocapable > The Vocapable Master Services Agreement: the core agreement governing platform use, incorporating the addenda, AUP, and AI disclosure. URL: https://www.vocapable.com/legal/terms ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. ## Incorporated into this Agreement These documents form part of the MSA. Where they conflict, the Telecom Compliance Addendum controls first, then the Platform Test Calls Addendum, then the AUP. - Telecom Compliance Addendum: Your calling-law obligations for consent, scrubbing, disclosures, and revocation. Controls over the MSA body where they conflict. - Platform Test Calls Addendum: The only permitted use of platform-owned numbers: verified test calls to numbers you control. - Acceptable Use Policy: What may never run on the Service, and what happens when it does. - AI Technology Disclosure: A plain-language notice of what our AI is, what can go wrong, and what relying on it means. - Data Processing Addendum: CCPA/CPRA service-provider terms for the personal information we process on your behalf, plus the two retention carve-outs that survive a deletion request. - Privacy Policy: how we handle personal information, incorporated by reference. # Vocapable Master Services Agreement Version: v1.3 · Effective date: August 15, 2026 This Master Services Agreement (this “Agreement”) is entered into between Capstra Labs, LLC, a California limited liability company doing business as Vocapable(“Vocapable,” “we,” “us”), and the entity accepting this Agreement (“Customer,” “you”). It governs Customer’s access to and use of Vocapable’s AI-voice calling platform and related services (the “Service”). By clicking to accept, by executing an order form referencing this Agreement, or by using the Service, Customer agrees to this Agreement. The individual accepting represents that they have authority to bind Customer; if not, Customer may not use the Service. The following documents are incorporated by reference and form part of this Agreement: the Telecom Compliance Addendum, the Platform Test Calls Addendum, theAcceptable Use Policy (“AUP”), the AI Technology Disclosure, the Data Processing Addendum (“DPA”), and the Privacy Policy. The Cold Outreach & Data Provenance Waiver is incorporated on and from Customer’s execution of it, and only with respect to the Elected Outreach it covers; a Customer that has not executed it is not bound by it and may not conduct Elected Outreach. If they conflict, the order of precedence is: the Cold Outreach & Data Provenance Waiver (as to Elected Outreach only, and only where executed), Telecom Compliance Addendum, Platform Test Calls Addendum, AUP, this Agreement’s body, then the remaining documents — except that the DPA controls over this Agreement’s body with respect to the processing of personal information, per its own Section 6. ## 1. The Service 1.1 What the Service is. The Service is a technology platform on which Customer configures, tests, and operates AI voice agents that place and receive telephone calls and send related messages, together with associated campaign management, compliance tooling, analytics, and APIs. 1.2 What the Service is not. Vocapable is a technology provider. Customer, and not Vocapable, determines who is called, when they are called, and what its agents are configured to say. Vocapable does not design, approve, or optimize Customer’s campaigns, calling lists, targeting, or scripts; does not source, sell, or recommend contact data; and does not act as Customer’s agent, telemarketer, or broadcaster. All production calls are placed on Customer’s own carrier account (Section 6) on Customer’s behalf and at Customer’s direction. 1.3 Compliance features are not legal advice. The Service includes compliance-support features (scrubbing, consent gating, calling-window enforcement, disclosure playback, opt-out handling, evidence records). These features assist Customer’s own compliance program; they are not a representation that any call is lawful, and no feature of the Service relieves Customer of its own obligation to comply with applicable law. Vocapable does not provide legal advice. 1.4 Modifications; no service level. The Service evolves. Vocapable may modify features, models, and interfaces, and will not materially degrade the core functionality of a paid subscription during its term without notice. The Service is offered without any service-level agreement or availability commitment. Availability targets, credits, or remedies for downtime exist only where expressly stated in a signed Order; otherwise, maintenance, degradation, and downtime — scheduled or unscheduled — are not breaches of this Agreement and create no right to any credit, refund, or set-off. ## 2. Accounts, Users, and Flow-Down 2.1 Customer is responsible for its accounts, credentials, and API keys, and for all activity under its account, including by its employees, contractors, and end users (“Users”). Acts and omissions of Users and of Customer’s affiliates using the Service are deemed Customer’s. 2.2 Downstream flow-down. If Customer uses the Service on behalf of, or resells to, its own clients, Customer must bind each such client to written terms no less protective of Vocapable than this Agreement (including the Telecom Compliance Addendum and AUP) and remains fully responsible for their compliance and conduct. 2.3 Eligibility and review. The Service is for business use only. Vocapable may require identity and business verification, may decline or limit any account, and may impose dial caps or other limits at its discretion, particularly during onboarding. ## 3. Fees, Usage, and Payment 3.1 Fees are as stated in the applicable order form or published pricing, plus metered usage. Usage is measured by the Service’s usage-event ledger, which is the system of record. Fees are exclusive of taxes; Customer pays all applicable taxes other than taxes on Vocapable’s income. 3.2 Customer pays for all usage. Unless Customer has configured limits (spend caps, concurrency caps, campaign limits), the Service does not stop traffic on Customer’s behalf, and Customer is responsible for payment of all usage incurred through its account, regardless of whether that usage exceeded Customer’s expectations, including usage caused by Customer’s own configuration, integrations, or Users. 3.3 Platform Test Calls are paid usage as described in the Platform Test Calls Addendum. 3.4 Undisputed amounts are due per the order form or on the payment method on file. Vocapable may suspend the Service for non-payment after 10 days’ notice. 3.5 No refunds; committed terms are non-cancellable. Except as expressly stated in Sections 12.2 and 12.3 or where required by non-waivable applicable law, all fees are non-refundable, and no refund or credit is provided for partial periods, unused capacity, unused included quantities, suspension caused by Customer, or dissatisfaction with the Service or its Outputs. The Sandbox account fee is non-refundable in all circumstances; its sole benefit beyond the Sandbox itself is the credit described in the Production Services Order. A committed term (the 3-year committed term) is a binding purchase of the full term, governed by the separately executed 3-Year Commitment Addendum, which controls with respect to the committed term: all committed fees are final, the term is non-cancellable for convenience, and if Customer abandons or terminates the committed term before it ends (other than under Section 12.2 for Vocapable’s uncured material breach), or Vocapable terminates it for Customer’s breach,the remaining fees committed for the term become immediately due and payable. Early termination does not relieve payment. 3.6 Chargebacks. Initiating a payment-card dispute or chargeback for fees validly incurred under this Agreement, instead of the dispute process in Section 13, is a material breach. Vocapable may suspend the Service immediately upon such a dispute, the disputed amounts remain due, and Customer is responsible for Vocapable’s reasonable costs of responding, including dispute fees charged by the payment processor. 3.7 Renewals; price changes. Month-to-month subscriptions renew each month until either party cancels, effective at the end of the then-current period; amounts already paid for the current period are not refunded. Committed terms renew, expire, or convert as stated in the applicable Order. Vocapable may change published pricing on at least 30 days’ notice, effective at Customer’s next renewal — never mid-term for a committed term. ## 4. AI Technology; Experimental Nature; Customer’s Evaluation Duty 4.1 Nature of the technology. Customer acknowledges that the Service uses artificial intelligence and machine learning technologies, including large language models and speech synthesis, that are probabilistic and non-deterministic. Outputs — including everything an AI voice agent says on a call, transcriptions, summaries, extracted data, analyses, and scores (collectively, “Outputs”) — are generated by automated systems and may contain errors, inaccuracies, omissions, fabrications (“hallucinations”), or unintended content; may be false, incomplete, misleading, offensive, or not reflective of recent events; may deviate from Customer’s configuration, instructions, prompts, or knowledge sources; and may differ across identical inputs. The Service is an evolving and experimental technology under ongoing development, and Vocapable makes no representation that its behavior will be consistent, repeatable, or free from error, before, during, or after any call. 4.2 Customer’s evaluation and review duty. It is Customer’s sole responsibility to evaluate whether the Service and its Outputs are appropriate for Customer’s use case, including determining where human review, verification, or supervision is appropriate.Customer will not rely on any factual assertion in an Output without independently verifying it. Customer is solely responsible for the personas, prompts, scripts, objectives, knowledge sources, guardrail configuration, and voice selections it deploys, and for all statements made by AI voice agents operating under its account. 4.3 No high-risk sole reliance. Customer will not use the Service, and will not permit its Users to use the Service, as the sole or determinative basis for any decision or communication that has a legal, financial, medical, safety-of-life, employment, insurance, housing, or similarly significant effect on any individual, without meaningful human review. 4.4 Notice to called parties. Customer is responsible for ensuring called parties receive all notices required by applicable law regarding the automated and AI-generated nature of calls. The Service plays a non-removable AI-disclosure opener on every call as a platform control; that control supplements, and does not replace, Customer’s own disclosure obligations. 4.5 Adverse outcomes. Customer acknowledges that errors of the kind described in Section 4.1 — including errors occurring before a call (e.g., in planning, scheduling, or list handling), during a call (e.g., misstatements by an agent), or after a call (e.g., in transcripts, summaries, or extracted data) — may negatively affect Customer or third parties, and that, as between the parties and to the maximum extent permitted by law, Customer bears that risk (subject to Sections 9–11). ## 5. Third-Party Providers 5.1 The Service interoperates with third-party providers, including large-language-model providers, model aggregators and routers, speech-to-text and text-to-speech vendors, telecommunications carriers, and payment processors (“Providers”). Providers are not part of the Service, and Vocapable is not responsible for them. Vocapable makes no representation or warranty on any Provider’s behalf and is not the agent of any Provider. 5.2 While Vocapable expects Providers to comport with industry norms and their own policies, Vocapable does not monitor and makes no guarantee or warranty in respect of any Provider’s compliance with its contractual or legal obligations, and shall have no liability for any Provider’s delay, degradation, outage, error, rate limiting, content filtering, model change or deprecation, discontinuation, or failure, or for any Output produced by a Provider’s model. 5.3 Bring-your-own credentials. Where Customer connects its own Provider accounts or credentials (including Customer’s own Twilio account or subaccount under Section 6), Customer’s use of those capabilities is governed by Customer’s own agreement with that Provider, and Customer is solely responsible for procuring, maintaining, funding, and complying with that agreement, including the Provider’s acceptable-use, voice, and messaging policies. ## 6. Telephony 6.1 Production traffic is Customer-carried. Except for Platform Test Calls, all voice traffic originated through the Service is transmitted using Customer’s own carrier account or subaccount, under Customer’s own carrier relationship, from telephone numbers procured and controlled by Customer. Customer is the originating party and the party on whose behalf all such calls are placed. Customer is responsible for its carrier registrations and obligations, including Robocall Mitigation Database registration where applicable, STIR/SHAKEN attestation, caller-ID authority, number reputation, and responding to industry traceback and regulatory inquiries within the timeframe stated in the Telecom Compliance Addendum. 6.2 Platform Test Calls. Telephone numbers under Vocapable’s own carrier account are available solely for paid, low-volume test calls as described in the Platform Test Calls Addendum. The Service enforces this restriction technically; any attempt to place other traffic on platform numbers is refused and is a material breach. 6.3 NO EMERGENCY SERVICES. THE SERVICE IS A SOFTWARE PLATFORM, NOT A TELEPHONE SERVICE, AND DOES NOT SUPPORT AND MUST NOT BE USED FOR EMERGENCY CALLS OR COMMUNICATIONS OF ANY KIND, INCLUDING CALLS TO 911, E911, 988, 112, OR ANY OTHER EMERGENCY OR SAFETY-OF-LIFE NUMBER OR SERVICE. CUSTOMER WILL NOT CONFIGURE ANY AGENT TO PLACE, RECEIVE, OR HANDLE EMERGENCY COMMUNICATIONS, WILL MAINTAIN ORDINARY TELEPHONE SERVICE FOR EMERGENCY USE, AND WILL SO INFORM ITS USERS. CALLS TO EMERGENCY NUMBERS ARE PROHIBITED AND ARE BLOCKED BY THE SERVICE WHERE TECHNICALLY FEASIBLE. ## 7. Customer Content; Data 7.1 Customer retains ownership of the contact lists, prompts, scripts, knowledge sources, recordings of its own personnel, and other materials it provides (“Customer Content”). Customer grants Vocapable the rights needed to operate the Service, including processing Customer Content and call data per the DPA. 7.2 Customer represents that it has all rights and lawful bases needed for Vocapable to process Customer Content as contemplated, including with respect to every contact record it imports. 7.3 Vocapable owns the Service, its models, software, and all improvements. Vocapable does not use Customer Content or call recordings to train foundation models. Usage data and telemetry, de-identified or aggregated so that no customer or called party is identifiable, may be used to operate, secure, benchmark, and improve the Service, consistent with the DPA. 7.4 Compliance evidence retention. Certain records (consent evidence, scrub results, disclosure and opt-out evidence, attestations) are retained per the DPA’s compliance carve-outs notwithstanding deletion requests, for at least five (5) years, because they are the parties’ legal-defense record. ## 8. Beta and Experimental Features Features identified as beta, preview, evaluation, trial, labs, or experimental are provided “as is” without any warranties or commitments of any kind, may be changed, suspended, or discontinued at any time without notice, may not be subject to the same security, availability, or support commitments as generally available features, andVocapable shall have no liability or indemnification obligation arising from or relating to Customer’s use of them. ## 9. Warranty Disclaimer EXCEPT AS EXPRESSLY STATED IN THIS AGREEMENT, THE SERVICE AND ALL OUTPUTS ARE PROVIDED“AS IS” AND “AS AVAILABLE,” WITH ALL FAULTS. VOCAPABLE DISCLAIMS ALL WARRANTIES, EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, INCLUDING MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, NON-INFRINGEMENT, AND ANY WARRANTY ARISING FROM COURSE OF DEALING OR TRADE USAGE. WITHOUT LIMITING THE FOREGOING, VOCAPABLE DOES NOT WARRANT THAT THE SERVICE WILL BE UNINTERRUPTED, AVAILABLE AT ANY PARTICULAR TIME, SECURE, OR ERROR-FREE; THAT OUTPUTS WILL BE ACCURATE, COMPLETE, RELIABLE, OR SUITABLE FOR ANY PURPOSE; THAT CALLS OR MESSAGES WILL BE DELIVERED, COMPLETED, OR RECORDED; OR ANYTHING WITH RESPECT TO THIRD-PARTY PROVIDERS. VOCAPABLE PROVIDES NO SERVICE-LEVEL COMMITMENT (SECTION 1.4). CUSTOMER’S USE OF THE SERVICE IS AT ITS SOLE RISK. ## 10. Indemnification 10.1 By Customer. To the fullest extent permitted by law, Customer will indemnify, defend, and hold harmless Vocapable, its members, managers, officers, employees, and agents from and against any and all claims, demands, actions, investigations, and proceedings (including by any government or regulatory authority), and all resulting losses, liabilities, civil penalties, statutory damages, regulatory fines and forfeitures, damages, settlements, costs, and expenses (including reasonable attorneys’ fees), arising out of or relating in any way to: (a) Customer’s or its Users’ access to or use of the Service; (b) calls placed, messages sent, or content spoken through Customer’s account, including any claim that such calls or messages violated the Telephone Consumer Protection Act (47 U.S.C. § 227) or FCC rules thereunder, the Telemarketing Sales Rule, the Telemarketing and Consumer Fraud and Abuse Prevention Act, the Truth in Caller ID Act, or any federal or state telemarketing, automated-calling, artificial-voice, AI-disclosure, call-recording, privacy, or do-not-call law; (c) Customer’s breach of this Agreement, the Telecom Compliance Addendum, the Platform Test Calls Addendum, or the AUP; (d) Customer’s failure to obtain, document, retain, or honor the revocation of consent; (e) Customer Content, prompts, scripts, knowledge sources, and calling lists; (f) any claim by a called party, by Customer’s own clients, or by any User; and (g) Customer’s violation of any third party’s rights. Vocapable may participate in the defense with its own counsel at its own expense; Customer will not settle any claim imposing non-monetary obligations on Vocapable without consent. 10.2 By Vocapable. Vocapable will defend Customer against third-party claims that the Service, as provided by Vocapable and used per this Agreement, infringes a U.S. patent, copyright, or trademark, or misappropriates a trade secret, and will pay resulting damages finally awarded or agreed in settlement. This obligation does not apply to claims arising from Customer Content or Outputs; combination with items not provided by Vocapable; modifications not made by Vocapable; use in violation of this Agreement; or Provider models or services. If the Service is enjoined, Vocapable may procure the right to continue, modify, replace, or terminate with a pro-rata refund. This Section 10.2 states Vocapable’s entire liability for infringement. 10.3 Insurance. An indemnity is worth what the indemnitor can pay. Accordingly, if Customer places more than 25,000 calls through the Service in any calendar month, then from the first day of the following calendar month and for as long as Customer continues to use the Service for production calling, Customer will maintain, at its own expense, with insurers rated A- or better: - Commercial general liability of at least $1,000,000 per occurrence and$2,000,000 aggregate; and - Technology errors & omissions / cyber liability of at least $1,000,000 per claim that does not exclude claims arising from telephone, telemarketing, or automated or artificial-voice calling, including claims under the Telephone Consumer Protection Act, and that covers statutory damages and regulatory proceedings to the extent insurable. Customer will name Vocapable as an additional insured on the general liability policy with respect to Customer’s use of the Service, will provide a certificate of insurance within ten (10) business days of written request, and will give thirty (30) days’ notice of cancellation or material reduction. Customer’s insurance is primary and non-contributory as to Vocapable for claims within Section 10.1. Maintaining insurance does not cap, reduce, or substitute for Customer’s obligations under Section 10.1, and a failure to maintain it is a material breach. ## 11. Limitation of Liability 11.1 Exclusion. TO THE MAXIMUM EXTENT PERMITTED BY LAW, NEITHER PARTY WILL BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, OR FOR LOST PROFITS, REVENUE, GOODWILL, DATA, OR COST OF SUBSTITUTE SERVICES, HOWEVER CAUSED AND UNDER ANY THEORY, EVEN IF ADVISED OF THE POSSIBILITY. 11.2 Cap. TO THE MAXIMUM EXTENT PERMITTED BY LAW, VOCAPABLE’S TOTAL AGGREGATE LIABILITY ARISING OUT OF OR RELATING TO THIS AGREEMENT OR THE SERVICE WILL NOT EXCEEDTHE GREATER OF (A) THE FEES PAID BY CUSTOMER FOR THE SERVICE IN THE TWELVE (12) MONTHS PRECEDING THE EVENT GIVING RISE TO LIABILITY, OR (B) ONE HUNDRED DOLLARS ($100). 11.3 Exclusions from the cap. Sections 11.1 and 11.2 do not limit: Customer’s indemnification obligations under Section 10.1; Customer’s breach of the Telecom Compliance Addendum, Platform Test Calls Addendum, or AUP; Customer’s payment obligations; or either party’s liability for fraud, willful injury to person or property, gross negligence, or willful violation of law, or any other liability that cannot lawfully be limited or excluded (including under California Civil Code § 1668). These carve-outs exist so that every other limitation in this Section is enforceable to its full stated extent. 11.4 The parties acknowledge these allocations reflect the pricing of the Service and are an essential basis of the bargain, and apply even if a remedy fails of its essential purpose. ## 12. Suspension; Termination 12.1 Vocapable may monitor, audit, meter, throttle, block, suspend, or terminate any campaign, agent, telephone number, feature, or account that Vocapable reasonably believes violates this Agreement, the addenda, the AUP, or applicable law, or that generates abnormal call patterns (including high unanswered-attempt volume, abnormally short average call duration, or elevated complaint, spam-label, or traceback rates), without prior notice where Vocapable reasonably deems it necessary to prevent harm, comply with law or a carrier or regulator requirement, or protect the platform. Vocapable will notify Customer as soon as practicable and lift the suspension when the basis is resolved. Vocapable has no liability for actions taken in good faith under this Section, and fees remain payable during suspension caused by Customer. 12.2 Either party may terminate for material breach uncured 30 days after notice (immediately for AUP, Telecom Compliance Addendum, or Platform Test Calls Addendum breaches), or for insolvency. If Customer terminates for Vocapable’s uncured material breach, Vocapable will refund the pro-rata unused portion of any prepaid fees for the terminated period — Customer’s sole and exclusive remedy for such a termination, subject to Section 11. Customer may export its data per the DPA for 30 days after termination; compliance evidence is retained per Section 7.4. Sections that by nature survive (3.5–3.6, 4, 5, 6.3, 7.4, 8–11, 13–15) survive. 12.3 Termination for convenience by Vocapable. Vocapable may terminate any subscription for convenience on 30 days’ notice, in which case it will refund the pro-rata unused portion of prepaid fees. Together with Section 12.2, this is the only refund provided under this Agreement. ## 13. Dispute Resolution; Arbitration; Class Waiver 13.1 Informal resolution first. The parties will first attempt in good faith to resolve any dispute by written notice and 30 days of discussion. 13.2 Binding arbitration. Any dispute, claim, or controversy arising out of or relating to this Agreement or the Service that is not resolved informally will be resolved by final and binding arbitration administered by the American Arbitration Association under its Commercial Arbitration Rules, by a single arbitrator, seated in the county of Vocapable’s principal place of business in California, conducted in English. The arbitrator may award any relief a court could award to the individual party. Judgment on the award may be entered in any court of competent jurisdiction. Each party bears its own attorneys’ fees except where the applicable rules or law provide otherwise. The arbitrator, and not any court, has exclusive authority to resolve disputes about the interpretation, applicability, or enforceability of this Section 13, except that a court decides the enforceability of Section 13.4. 13.3 Exceptions. Either party may (a) bring an individual claim in small-claims court, and (b) seek temporary injunctive relief in court for intellectual-property infringement, misuse of confidential information, or to prevent unauthorized use of the Service, pending arbitration. 13.4 Class, collective, representative, and mass-action waiver. EACH PARTY MAY BRING CLAIMS AGAINST THE OTHER ONLY IN ITS INDIVIDUAL CAPACITY, AND NOT AS A PLAINTIFF OR MEMBER IN ANY PURPORTED CLASS, COLLECTIVE, CONSOLIDATED, REPRESENTATIVE, OR MASS ACTION OR ARBITRATION. The arbitrator may not consolidate more than one party’s claims or preside over any form of representative or class proceeding. 13.5 Blow-up severability. If the waiver in Section 13.4 is held unenforceable as to a particular claim, then that claim (and only that claim) shall proceed in court rather than in arbitration, and Section 13.6 applies; the waiver remains enforceable as to all other claims. Except as stated in this Section 13.5, if any part of this Section 13 is held unenforceable, the remainder remains in effect. 13.6 Governing law; venue. This Agreement is governed by the laws of the State of California, without regard to conflict-of-laws rules. For matters properly in court under this Section 13, the parties consent to the exclusive jurisdiction of the state and federal courts located in the county of Vocapable’s principal place of business in California, and waive any right to a jury trial to the maximum extent permitted by law. 13.7 Time to bring claims. To the maximum extent permitted by law, any claim arising out of or relating to this Agreement or the Service must be filed within one (1) year after the claim accrued, or it is permanently barred. This Section does not shorten any period that applicable law does not permit to be shortened. ## 14. Electronic Acceptance This Agreement and all incorporated documents may be accepted electronically. Customer consents to contracting electronically under the U.S. E-SIGN Act and California UETA. Vocapable records, for each acceptance: the document version, a cryptographic hash of the exact text presented, the accepting individual’s name, title, and email, the originating IP address, and a server-side timestamp. Those records are the parties’ authoritative evidence of acceptance. When Vocapable updates a document, continued use of the affected Service features requires acceptance of the updated version; the Service will prompt for re-acceptance and may restrict the affected features until accepted. ## 15. General Independent contractors; no third-party beneficiaries (except Vocapable indemnitees under 10.1). Customer may not assign without consent except to a successor in a merger or asset sale not involving a Vocapable competitor; Vocapable may assign to an affiliate or successor. Notices to support@capstralabs.com and Customer’s account email. Force majeure for events beyond reasonable control (including carrier and Provider failures). Export compliance and anti-corruption compliance. Entire agreement; order of precedence per the preamble; no reliance on any statement not in this Agreement. Amendments per Section 14. If any provision is unenforceable, it is modified to the minimum extent necessary and the remainder stands. No waiver by conduct. Headings are for convenience. Capstra Labs, LLC, a California limited liability company, d/b/a Vocapable · Questions: support@capstralabs.com --- # Platform Test Calls Addendum | Vocapable > Vocapable Platform Test Calls Addendum: the only permitted use of platform-owned numbers, which is verified test calls to numbers you control. URL: https://www.vocapable.com/legal/test-calls-addendum ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Platform Test Calls Addendum Version: v1.0 · Incorporated into the Vocapable Master Services Agreement. This Addendum governs the only use of telephone numbers held under Vocapable’s own carrier account (“Platform Numbers”): letting Customer hear its own AI agent on a real phone call before that agent ever reaches a real contact. ## 1. What Platform Test Calls Are 1.1 A “Platform Test Call” is a call placed from a Platform Number solely to a telephone number that Customer has verified through the Service’s number-verification process (a one-time code delivered to that number and confirmed by Customer), demonstrating that Customer owns or controls the destination or that its holder has expressly consented to receive test calls. 1.2 Platform Test Calls are paid usage, metered and billed like other usage, and arevolume-capped, monitored, and logged. Vocapable may adjust caps, meter rates, and availability at any time. ## 2. What Platform Numbers May Never Carry 2.1 Platform Numbers may not be used for any production, marketing, sales, outreach, survey, collection, reminder, notification, or campaign traffic, or for any call to any person other than a verified destination under Section 1.1. The Service enforces this technically: a call from a Platform Number to a non-verified destination is refused by the dialing gate. Circumventing or attempting to circumvent this restriction is a material breach permitting immediate termination. 2.2 Customer represents and warrants, for each verified destination, that Customer owns or controls the number or has the express consent of its holder to receive test calls, and that no verified destination will be used to reach any consumer or third party for any commercial purpose. ## 3. Production Traffic 3.1 All other traffic rides Customer’s own carrier account under MSA §6 and the Telecom Compliance Addendum. Vocapable may make managed production telephony available in the future under separate terms; until then, no such service exists regardless of any statement outside the Agreement. ## 4. Suspension 4.1 Vocapable may suspend Platform Test Calls at any time, with or without notice, to protect number reputation, comply with carrier requirements, or investigate suspected misuse, without liability. --- # Website Terms of Use | Vocapable > Vocapable Website Terms of Use: terms for this website only, because platform use is governed by the Master Services Agreement. URL: https://www.vocapable.com/legal/website-terms ## All legal documents - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use - Trust Center The website and the product serve the same source text for every document, so the published text and any accepted text can never drift. # Website Terms of Use Version: v1.0 · Effective: August 2, 2026 These terms govern use of vocapable.com and its subdomains (the “Site”), operated by Capstra Labs, LLC d/b/a Vocapable. Use of the Vocapable platform itself is governed by the Master Services Agreement, not these terms. - Content. Site content is provided for general information, “as is,” without warranties. Product screenshots show real product surfaces with synthetic data. Nothing on the Site is legal advice, a service commitment, or an offer capable of acceptance; the platform is sold only under the MSA. - No compliance representation. Descriptions of compliance features describe platform functionality. They are not a representation that any use of the platform is lawful for you, and they do not create any advisory relationship. - IP. The Site and its content are ours or our licensors’. Limited license to view; no scraping, framing, or use of our marks without permission. - Acceptable use. No interfering with the Site, probing security, or automated bulk collection. - Third-party links are provided for convenience; we are not responsible for them. - Liability. To the maximum extent permitted by law, we are not liable for damages arising from Site use; total liability for Site claims is capped at $100. - Governing law. California; exclusive venue in the state and federal courts of the county of our principal place of business in California. - Changes. We may update these terms; continued use after posting is acceptance. Contact: support@capstralabs.com --- # Outcomes: the metrics we report | Vocapable > Who picked up, who qualified, what each outcome cost. Every metric Vocapable reports is computed on your own campaigns, with the /v1 read behind it named. URL: https://www.vocapable.com/outcomes Outcomes · What we measure # What each conversation cost, while the campaign is still running Who picked up. Who qualified. What every outcome cost you. Each figure is computed on your own campaigns, off ledgers the platform already holds. Each metric is a live API report; supported CSV exports provide the underlying record types for downstream analysis. For enabled events, call completion reaches your signed webhook, while available call evidence stays behind the authenticated /v1 read. ## The reports, and the grammar they share The funnel, the disposition mix and the turn timings are reads on the same API your dashboards use. Each one takes a window, an optional grouping (by campaign, by published agent version, or by day) and an optional filter down to a single campaign. Same shape whichever one you ask for, so a query you write once keeps working. ### The funnel GET /v1/analytics/funnel Attempts, dials, connects, conversations, outcomes and goal completions, plus four rates computed from them: connects over dials, conversations over connects, outcomes over attempts, and goal completions over conversations. ### The disposition mix GET /v1/analytics/dispositions Every attempt lands on one value from a pinned vocabulary: completed, answered but incomplete, voicemail, no answer, opted out, refused by policy, and the rest. The report returns each value by name with its share of the group it sits in, so the mix reads the same way next quarter. ### Turn timings GET /v1/analytics/latency Turn-level timing percentiles read off your own calls. This one needs the call-read scope as well as the campaign one, because a turn ledger is a transcript and is treated like one. Ask for a window outside the bounds and you get a problem document naming the rule you hit, so you fix the query instead of squinting at a half-drawn chart. Longer histories leave through CSV export jobs: supported record types, the filters and columns you choose, and a signed download while the completed job is retained. A stage with nothing beneath it reports empty, never zero. A campaign that dialed nobody has no contact rate, so your chart never carries a point no call produced. The rule lives in the code that computes the rate rather than in a dashboard's formatting, which is why it holds on the API read and on the screen. ## What each scenario reports Bare names are cheap, so every one below carries the sentence that says what it counts. If a name cannot take that sentence, it does not belong on the panel. ### Speed-to-lead - Contact rate How many of the leads you chose to call actually picked up and talked. - Qualification rate How many conversations ended qualified against your own rubric, per campaign. - Disposition mix Every attempt written from a pinned vocabulary, so the mix reads the same tomorrow. - Refusals, by reason The dials the platform refused to place, each with the value that caused it. - Demos requested Interested leads whose call captured the window they named, in their own words. - Cost per qualified outcome Spend to date over the outcomes the same payload counted. Division, not a model. ### Staffing calls - Contact rate How many of the candidates you chose to call actually picked up and talked. - Confirmed shifts and decline reasons Each response is saved as a structured shift-confirmation outcome, so the team can work confirmed coverage and clear decline reasons without replaying every call. - Cost per completed call Read off the usage ledger for the exact calls this campaign made. ### Feedback calls - Contact rate How many of the customers you chose to ask actually picked up and talked. - Completed conversations Feedback calls that reached the end of your questions, counted per campaign. - Survey responses and scores Each recorded response is saved as a structured survey outcome, so feedback teams can work answers, scores, and CSAT/NPS rollups from the program they ran. - Cost per completed call Read off the usage ledger for the exact calls this campaign made. A name appears here once a read produces it, which is why every metric on this page works on the first campaign you run. The one nobody else reports ## Refusals, by reason Every dial the platform held back is counted and named, each beside the value that caused it. It is the row that tells you a campaign stayed inside your own rules while it was still running, instead of six weeks later when somebody asks. It is computed twice, on both sides of the dial. The batch run writes a breakdown of what it suppressed before the campaign launches, sorted and rendered in your own run report. The dial-time re-check writes the refusal onto the attempt with its sub-code, and the disposition report returns it by name with its share. What the platform performs, named one by one: phone normalization, suppression against your own internal do-not-call list at the batch run and again at dial time, and the dialability gate. Those three are what a refusal count is made of. The breakdown shows the reasons that actually fired on your calls, each with the count it took off the list. A held-back dial still writes its file A dial the platform holds back is written down as carefully as one that connects: the checks that ran, the value that stopped it, and the row it took off the list, exportable with the rest of the run. The answer to “why was this number never called” is on file the day somebody asks for it, not reconstructed the week after. The whole shape of a call’s record is on the Trust Center. ## Cost sits on the same counter as the outcome Every campaign carries its spend to date beside its dials, its connects and its goal completions, on one payload your dashboards poll and your own code can read. The cost of a completed call is division over two figures you already hold, so you can check our arithmetic in a spreadsheet. Why the figure is this small. It counts the platform’s own charges. Production calling runs on your own Twilio account, so your carrier bills you for the minutes at the rate you already negotiated, and a Vocapable invoice never carries a carrier line for you to reconcile. ### Cost per completed call Read off the usage ledger for the exact calls this campaign made. ### Spend to date On the live campaign counter, beside the dials and the goal completions it is divided by. ### Usage detail Event-level, per meter, on a read of its own, so you can reconcile the total rather than trust it. The published rates behind every one of those figures are on the pricing page, and every field named here is a /v1 read before it is a tile. ## Call completion arrives first; analysis follows For the events you enable, the platform posts a signedcall.ended event at call completion: the disposition, duration and end reason, the identifiers that tie it to the contact and campaign, and whether a transcript or recording is available. It marks post-call analysis as pending. Fetch available call evidence through authenticated/v1 reads; a latercall.analyzed event carries the analysis result. Delivery retries when your configured endpoint is down, so a rep can work the next step in their own tooling while the conversation is fresh. Your stack stays the system of record. For enabled events, shift confirmations emitshift.confirmed, and survey responses emitsurvey.completed; each is also its own/v1 resource. CSV export jobs cover supported record types with the filters and columns you choose, then provide a signed, time-bounded download when the job is complete. ### Call evidence, when available Fetch ordered turns, transcripts, or recordings through authenticated/v1 call reads. ### One value, pinned The disposition comes from a fixed vocabulary, so your reports keep their meaning. ### Structured results, ready to use Survey answers and scores, or a shift confirmation, are saved as their own resources and delivered through their corresponding configured signed events. The trust close ## Numbers you can produce, on calls you placed You choose who gets called, and you and your advisors set the standard that choice has to meet. From there the platform does the work: it enforces the calling window, the frequency caps, the consent standard and the recording-consent class in force on every attempt, stricter-only, and it writes down which values bound the call. Every call the platform places, and every one it holds back, writes an evidence-bearing call or attempt record: which checks ran, which were waived, and on whose signature. The corresponding authenticated /v1 read and supported CSV export job keep that evidence ready when somebody asks how a call happened. The whole claim The reports above read the same rows that evidence record does, so any figure on this panel walks straight back to the individual calls that produced it: in the contact record, in a CSV job for its supported record type, or on a /v1 read. One number, one set of calls, one place to check it. ## Questions people ask first From your own campaigns and nowhere else. Each figure is read off a ledger the platform already holds: the attempt row, the call record, the usage meter. The same /v1 reads that feed your dashboards return them, so every number on the panel is one your own calls produced and every one of them walks back to the individual calls behind it. It reports empty rather than zero. A campaign that dialed nobody has no contact rate, so the chart stays honest instead of drawing a point your calls never produced. The rule lives in the code that computes the rate, which is why it holds on the API read and dashboard view. Your own calendar holds that fact, and your team is the one who marks it. What the panel reports is everything leading up to it: the conversation itself, the value the attempt was disposed with, the window the prospect named in their own words, and the record your team can retrieve through its configured delivery path. The dials the platform held back before they were placed: your internal do-not-call list applied at the batch run and again at dial time, the dialability gate, and the calling window and frequency caps in force. Each one arrives named, with the value that caused it, so you can see a campaign staying inside your own rules while it is still running rather than six weeks later. Yes. Every report on this page is a /v1 read, and the dashboard is one consumer of it rather than the source. Longer histories leave through CSV export jobs for supported record types; a completed job provides a signed download for its time-bounded window. For the events you enable, call completion arrives as a signed call.ended event with post-call analysis marked pending; call.analyzed arrives when that analysis is ready. Structured survey responses and shift confirmations use their own enabled signed events. ## Run a campaign and read your own panel The sandbox runs the whole dial path with every gate live and failing closed, and it writes the same rows these reports read. The first numbers you see here will be yours. --- # Press kit | Vocapable > The Vocapable press kit: approved boilerplate, the entity facts behind Capstra Labs, LLC, brand files, and what we will confirm on the record. URL: https://www.vocapable.com/press Company · Press kit # Everything you need to describe us correctly The entity behind the product, the wording we approve, the brand files, and what we will confirm on the record. Everything here is ready to paste into a draft today. Write to support@capstralabs.com and a person answers. ## The approved wording Copy this as it stands. It is authored once in the site’s own source and read by every surface that needs it, which is why a paraphrase drifts and this paragraph does not. Boilerplate Vocapable is a product of Capstra Labs, LLC: AI voice agents that place and answer business phone calls on the customer’s own Twilio account and their own numbers. Every call returns a record, and the compliance controls behind it cannot be switched off. One line, for a caption or a listing AI voice agents that qualify demand, run workforce and feedback calls, and answer inbound callers on your own Twilio account. Every conversation becomes usable work. If you name the stack Vocapable runs on Twilio; no partnership or endorsement is implied. Worth a sentence of its own, because the shorthand a writer reaches for gets the economics backwards. Twilio is infrastructure our customers hold accounts on and our software speaks to. Every production call rides the customer’s own account at the carrier rate they already negotiated, so a Vocapable invoice prices the AI session and never the carriage. On first mention the company is Capstra Labs, LLC, doing business as Vocapable. After that, Vocapable on its own is correct. ## The facts Each of these is checkable against a document published on this same site, which is the only kind of fact worth putting in a kit. The whole company page is at /about, every agreement is published in full at /legal, and the security and compliance posture is on the Trust Center. ## Brand The mark is four bars of different heights on a rounded tile — a level meter, because the product is a voice on a phone call. Every file below is published and servable right now. Shown from favicon.svg, which is the vector original. The square mark as a raster tile, at the largest size published. The default choice when a listing asks for one file. The same mark with the safe padding a maskable icon needs, so a platform that crops to a circle does not cut into the bars. The compact raster tile, for a directory row or a card that renders the mark small. The home-screen tile iOS asks for, published so an add-to-home-screen does not fall back to a screenshot. The mark as vector artwork. Reach for this one for print, or for any size the rasters cannot hold. The browser-tab icon, kept for surfaces that still ask for this format specifically. ### Colours ### The name beside the mark Our own header sets Vocapable in type beside the mark rather than shipping a drawn lockup, so you can do the same in one step: set the name in your own type beside any icon file above and you have exactly what the header has, at exactly the size your layout wants. ### Using them The mark and the name Vocapable belong to Capstra Labs, LLC. Use them to refer to us and keep the proportions and the colours as supplied. Any claim of partnership or endorsement needs a signed agreement behind it. ## Who to quote Joseph Opanel Founder Interview requests go to the same address: support@capstralabs.com ## Where to hear it The published demo calls are scripted simulations in AI voices, labelled as simulations on the player, and every one prints its transcript beside it. The sample evidence file shows what the platform records for a single call attempt. Both are open pages: no form, no email address, nothing to request. They are on /resources, with the evidence file at /evidence. ## What we confirm The entity facts on this page, the published rate card, and anything in the agreement set. All three are public already, so a quote can be checked against the source before you file it. The rate card is at /pricing and it is the whole of it; the agreement set is at /legal, published rather than described. ## The fastest way to describe the product is to open it. The sandbox is free, and the whole builder is in it. --- # Pricing | Vocapable > Published plans and usage rates for Engage, Workforce, Sales and inbound Receptionist. Price your own campaign in the calculator. The sandbox is free. URL: https://www.vocapable.com/pricing # Pricing Every rate we bill is on this page. One monthly price per plan, $0.005 per call, and AI usage from $0.13/min, down to $0.07/min on enterprise volume. The calls leave your own Twilio account, so your Twilio rate is the rate you pay for minutes. Start free in the sandbox; $0 implementation fee. The compliance floor is on every plan at $0.The AI disclosure, opt-out detection, internal do-not-call suppression, phone normalization and the per-call evidence file ship on every class, enforced in code, at no charge. Design-partner pricing, published so you can plan, and fixed for the term of any order you sign. Step one · every plan starts here ## The sandbox is free The full builder, direct no-tool drafts, and unlimited simulated calls that are never metered: your script down the whole dial path, every compliance gate live, every disposition written the way production writes it. No credit card and no implementation fee. Nothing to connect, nothing to sign. This is where you find out whether you want the rest, before a dollar changes hands. ## Priced by what your calls carry One engine. Three outbound workload classes, priced by the compliance weight and capacity your calling actually uses, where a class is a ceiling rather than a channel: higher classes include the work of the ones below. Receptionist is the inbound half of the same engine. It answers your calls instead of placing them, and you can buy it on its own or add it to any outbound plan for the same monthly money. Every figure below is a rate we bill: the monthly plan price, the per-call fee and the AI-minute rate on either term. Carrier minutes are not in these numbers. Twilio bills those to you directly, at the rates you negotiated, which is a large part of why the numbers below are this small. ### Engage CSAT & surveys, transactional and relationship outreach $349 /mo - Feedback conversations in your voice - Verbatims on the contact record, exportable - Records and exports for every call - Full compliance floor at $0 ### Workforce Your own people: staffing, internal, opt-in non-sales $549 /mo - Staffing calls at roster scale - Answers and decline reasons, documented - Records delivered by webhook or CSV - Full compliance floor at $0 ### Sales Outbound revenue motions on your opt-in list $1,199 /mo - Qualification against your rubric - Qualified lead outcomes delivered to your configured webhook - Includes everything in Workforce and Engage - Full compliance floor at $0 ### Receptionist Inbound answering on your own numbers: menu, team, AI, voicemail $250 /mo - Your own answering steps, in your order - Voicemail recorded to the call record when nobody is free - Business hours and holidays in your own timezone - Full compliance floor at $0 The outbound plans bill four things: the monthly price with its included concurrent sessions, $200/mo for each additional concurrent session, $0.005 per call attempt, and AI usage at $0.13/min month-to-month or $0.09/min on a 3-year commitment, with the volume discount below stepping that minute rate down further. Receptionist bills $250/mo with its first inbound connection included and $150/mo for each simultaneous connection after it, and that is the whole of a Receptionist invoice: answered calls carry $0 in per-call fees and $0 in AI minutes, because both charges ride the calls your account places. Twilio bills your carrier service to you separately, on your own bill, at your own rate. One company, several jobs ## Reception, workforce and sales. Together. A service company needs to answer customers, support its workers, call the roster and qualify new customer applications. Use Sales for workforce and sales campaigns, then add inbound receptionist connections to the same subscription. - Customer and worker reception: configure the answering assistants for each audience. - Workforce outreach: capture availability and shift responses for your coordinators. - Sales qualification: ask about fit and interest using your company’s criteria. Inbound connections and outbound sessions are separate capacity pools. Adding an assistant does not add a simultaneous call slot. Choose capacity for the calls you need to handle at the same time. Example configuration $1,599/month fixed Plus outbound usage and your separate carrier bill. Handle up to 2 inbound calls and 2 outbound conversations at once across your assistants. Add outbound sessions at $200/mo each and further inbound connections at $150/mo each. Month-to-month outbound usage is $0.13 per AI minute and $0.005 per attempt. Answered inbound calls carry no Vocapable per-minute or per-call charge. This example shows recurring costs; an existing subscription’s upgrade payment is shown separately before purchase. ## How production works Build and validate the workflow before you pay. When you are ready, choose a plan, accept the production order, add a payment method, and connect your carrier in the portal. Ready workspaces activate there; a workspace needing a carrier or compliance decision carries its configuration into review. ### Sandbox Free, no credit card. Build your agent and hear the simulated demo calls. Your real script runs the whole dial path with every compliance gate live, before a number points anywhere. ### Choose your plan Choose a class and term, accept the Production Services Order with published rates, and add a payment method. The portal shows the next readiness step. Month to month bills AI usage at $0.13/min; a 3-year commitment bills it at $0.09/min on the same base, non-cancellable and all sales final under a signed non-refund waiver. ### Activate on your carrier Connect and verify your own Twilio subaccount in the portal: your numbers, registrations, and caller-ID reputation remain yours. Ready workspaces activate production there. If a carrier or compliance decision needs attention, the same readiness context routes to review. Usage rates and volume discount ## Enterprise plans as low as $0.07/min For every $10,000 of total monthly spend - plan base, session add-ons, call fees, and AI minutes together - the AI-minute rate steps down $0.01, to a floor of $0.07/min. A step applies mid-month, going forward from the moment your spend crosses the threshold, and the meter resets at the start of each billing month. The discount stacks with the 3-year committed rate, so both tables below end at the same floor. ### Month to month ### 3-year commitment You never have to ask for it. Cross a threshold and the next minute bills at the lower rate, in the same month. ## Price your own campaign shape Your list, your assumptions, our published rates. Every field starts on a placeholder rather than a benchmark of ours: put your own numbers in and watch the monthly total move. ### Your assumptions The days field paces the calling. It moves the attempts-per-day figure and nothing else: this campaign costs the same whether you work the list in a week or a month. Estimated monthly total $2,384 An estimate, not a quote, priced at the undiscounted published rate. The volume discount brackets on your whole account's realized monthly spend, which one campaign shape cannot know, so it can only move this figure down. Thetier tables show where it steps. Carrier minutes sit outside both figures. Calls leave your Twilio account and Twilio bills those minutes to you at your own rate. Only the contacts a pass missed are retried, so the attempts figure is the sum of each pass over whoever is left rather than every contact times every attempt. Reaching the same person on a later attempt is still one conversation. Usage is the conversations times your minutes times the AI minute rate, plus every attempt at $0.005. The plan base is your class plus any extra concurrent sessions at $200 each. Receptionist prices differently, because it answers calls rather than placing them: its invoice is the monthly price plus any additional connections, with no campaign shape to rate. How the receptionist works. Comparing this card against a per-minute quote from somewhere else? What that arithmetic leaves out ison its own page. ## Pricing questions, answered A sandbox account for your team, free, with no credit card. The full builder, direct no-tool drafts, and unlimited simulated calls that are never metered: your script down the whole dial path, every compliance gate live, every disposition written the way production writes it. You pay when you activate production, and not a day before. The sample calls on this site are simulated demos in AI voices, transcripts printed beside them, free and behind no form. Hearing your own agent is what the free sandbox opens: build it with your rubric, your material and your voice, then run it end to end against simulated telephony with every compliance gate live. The agent worth judging is the one you configured, and the account is free, needs no credit card, and does not expire. Choose your workload class and term, accept the Production Services Order with its rates in writing, add a payment method, and connect the carrier account you already own. The portal shows the readiness steps for your workspace; ready workspaces activate production there, while a carrier or compliance decision can route the same context to review. Calls run on your own verified Twilio subaccount, your numbers and your carrier bill. Terms are month to month at $0.13 per AI minute, or a 3-year commitment at $0.09 per AI minute on the same monthly base price. For production, yes, and it is the reason the per-minute number on this page is this small. Calls leave a Twilio subaccount you own: your numbers, your registrations, the rates you negotiated, and the caller-ID reputation those numbers earn. Twilio bills the carrier service straight to you at that rate. Your Vocapable bill is the plan, the call fee and the AI minutes, which is the whole of it. You connect the subaccount once in the portal and verify it, and it stays yours to take anywhere. Building in the sandbox needs none of it: you connect a carrier the day you are ready to dial someone. The classes price capacity and the calling controls each workload needs. A roster is bursty and deadline-bound, so Workforce reserves more concurrent sessions than Engage to work it in time; Sales adds the speed-to-lead motion on top. A class is a ceiling rather than a channel: a Sales subscription runs Workforce and Engage work freely, on the same subscription. Yes. Sales includes Workforce and Engage calling, and you can add receptionist connections to that same subscription. Customer and worker reception share your inbound capacity; workforce and sales campaigns share your outbound capacity. For example, Sales with 2 outbound sessions and 2 inbound connections has a fixed monthly price of $1,599, plus outbound usage and your separate carrier bill. Configure the assistants for their different jobs and add receptionist capacity in Billing. Concurrency is a line item you buy rather than a headcount you hire. Engage includes 1, Workforce includes 2, and Sales includes 2; additional concurrent sessions are $200 per month each, in any quantity, on any outbound plan. A concurrent session is a reserved calling slot: it sets how many conversations can run at the same moment, and it is platform capacity rather than a phone line. Receptionist buys answering rather than dialing, so its capacity is counted separately: $250 per month includes 1 concurrent inbound connection, with further simultaneous connections at $150 per month each. Inbound receptionist connections stay reserved separately from outbound campaign sessions. It drops your AI-usage rate from $0.13 a minute to $0.09 and holds it there for the term. The monthly base price, the $200 extra-session price and the $0.005 per-call fee are the same on both terms, and the volume discount still stacks on top, down to the same $0.07 per minute floor. Both sides sign it: a 3-year commitment is non-cancellable and all sales are final under a signed non-refund waiver. Only if you want the workspace. Send data by CSV or API and run everything from the dashboards, or run it headless: your stack stays the system of record, contacts are created through the API as you dial, and enabled outcome events can reach your configured webhook. Same engine, same rates, same compliance floor either way, and production runs on your own Twilio account in both. Sign an order and your rates are fixed for its term, which is the whole reason to sign one. The card itself is design-partner pricing and can move as we validate capacity on real calls. Every figure we bill is on this page, which is the point of publishing it: you can do the sum before you talk to us, and the sum you did is the sum you get. Four controls that no plan, flag, or support tool can switch off: the AI-disclosure opener played before the first model token, opt-out detection writing your internal do-not-call list, that list applied at the batch run and re-checked again at dial time, and phone normalization. The per-call evidence file comes with them, and so does enforcement of the calling window in the contact’s local time and the frequency caps. Every plan, every class, $0, on the cheapest seat and the largest order alike. ## The prices are public. The sandbox is free. Open one today, or talk to sales about enterprise volume and 3-year terms. --- # AI receptionist for your inbound calls | Vocapable > An AI receptionist on the number you already publish, answering on the Twilio account you already own. You write the steps it follows. From $250 a month. URL: https://www.vocapable.com/receptionist Product · The AI receptionist # Your number, configured around your business $250 a month puts an AI receptionist on the number you already publish, on the Twilio account you already own. You write what it does: a menu, your own reps, the agent, a voicemail. Answer fifty calls this month or five hundred, and the Vocapable price is the same. ## You write what happens, step by step One answering document per number, holding two ordered chains: one for your open hours, one for everything else. Your hours and your holiday dates, in your own timezone. Six kinds of step, in any order you like. - 1 A keypress menu Speak the options, read the key they press, follow that branch. Menus nest four levels deep, and you write the branch for the caller who presses nothing. - 2 Ring your own people Your reps, in the order you listed them. Destinations resolve at the moment the call is built, so the rep who changed their number this morning still gets the call. - 3 The AI agent Hands the caller to a published version of your agent, which answers in English from the material you gave it. This is the step people mean when they say receptionist. - 4 Voicemail Takes a message and keeps the audio on the call record with the rest of that call’s evidence. You press play. - 5 A spoken message Say one thing and hang up. It is the fastest closed-hours answer you can publish. - 6 Forward Bridge the caller straight to a number you chose. Your carrier carries it, and it claims none of your connections. You have room for sixty steps across both chains and menus four levels deep, and the editor holds you to that when you save. A caller never finds the edge of it. Business knowledge and caller requests ## Keep the receptionist informed, and keep follow-up visible Give the agent reviewed business facts, then turn the calls that need a person into work the right teammate can see. The knowledge and request workflow lives alongside the answering flow, not in a separate system. ### Reviewed business knowledge Build facts from documents or website source snapshots, review what belongs in the receptionist, and keep the business information your callers need in one place. ### Preview, then publish Preview the reviewed facts an agent version will use, then publish a version you can trace. Published versions stay tied to the knowledge and configuration they were released with. ### An Inbox for caller requests The agent can capture messages, open tickets, schedule callbacks, or update a known caller's lead. Your team can assign and resolve those requests from the operational Inbox. Release control ## Change the answer with evidence, not guesswork Author regression cases for the questions that matter to your business and keep release evaluation evidence with the version you publish. Each release has a traceable configuration, knowledge set, and proof of what was checked. ### A recoverable version history Published agent versions are immutable. When a compatible healthy version needs to return, the platform can restore it with the matching configuration rather than recreating the receptionist from scratch. ## Only one of those steps costs you a connection Menus, rings, forwards and spoken messages run on the carrier leg and claim none of your connections. That is why one connection can publish an answering flow far larger than the concurrency you bought: most of the flow costs you nothing to hold. The AI step is the one you are actually buying. It claims a reserved inbound connection for as long as the conversation lasts, and gives it back the moment the caller hangs up. What claims a connection - AI The agent conversation. One connection, for as long as it lasts. - MenuNone. - RingNone. - VoicemailNone. - SayNone. - ForwardNone. Your carrier still bills you for those seconds, at the rate you already pay it. What a connection actually buys ## Reserve inbound capacity for your receptionist A connection you buy is held for your inbound AI flow from the moment its charge clears, including while your own dialer is running at full pace. It is held whether or not anyone is on the line, so inbound and outbound capacity stay distinct. When a caller arrives while your connections are busy, the call keeps moving: it advances to the next step you wrote down, which is usually your own voicemail. What happens then is the thing you chose, which is why the last step in the chain is the one worth writing carefully. ## The whole price, on one line The plan $250/mo Includes 1 concurrent inbound connection on the number you already publish. Each additional connection $150/mo In any quantity. Each one lets one more caller talk to your agent at the same moment. On its own, or attached Receptionist answers the phone. Attach it to an Engage, Workforce or Sales subscription when you want the agent placing calls too, for the same monthly money. Inbound connections and outbound sessions never draw from one another. The plan plus your connections is the whole Vocapable invoice. The calls it answers add nothing per call and nothing per minute, so a busy month and a quiet one come to the same number. Carrier minutes stay where they already are: on your own Twilio bill, at the rate you negotiated. Adding a connection charges a full month for it up front, whatever day of the period you buy on, and the checkout shows you the exact amount and the exact date your subscription next renews before you agree. The line goes live when that charge clears. A connection you already hold keeps its grace period, so one failed card does not take your receptionist off the phone. The trust close ## Every answered call leaves a file you can hand to somebody Four controls run on a call you answer exactly as they run on a call you place, and no setting in your answering document can knock them out of place: the AI disclosure plays before the first model token, an opt-out is heard and written to your internal do-not-call list, that list suppresses, and numbers are normalized before anything reads them. The checks that belong to outbound calling are recorded with the reason they do not apply, so the file explains itself to whoever reads it next. A check that did not apply never reads as a check that passed. Recorded on an answered call - AI disclosure: played before the first model token. - Opt-out: heard, and written to your internal do-not-call list. - Suppression: that list applied before the next contact. - Numbers: normalized before anything reads them. - The controls that belong to outbound calling: written down with the reason they did not apply, because this caller dialed you. Read the whole shape of a call’s record on the Trust Center. ## Where a connection earns its money Three moments in an ordinary week, and the same one line covers all of them. ### Nine at night Your closed-hours chain picks up. A spoken message, a menu, or the agent taking the whole enquiry. In the morning the call record is waiting for you. ### While your reps are talking One caller is already on with your team. The next one reaches the agent, because ringing your own people claims none of your connections. ### The question you answer all day Hours, address, what you charge, what you do. The agent answers it off the material you gave it, in your words, and hands over anything you want a person for. ## What getting started actually looks like Three steps, and the middle one is your carrier’s. See pricing for the plan itself. ### 1 · Verify your Twilio account Your own subaccount, your own credentials, connected once. Start here: it is the step with the most waiting in it. ### 2 · Point your number at us The number you already publish, aimed at a signed webhook. Point it back whenever you like. It stayed yours the whole time. ### 3 · Write the steps, buy the line Compose the whole answering document first, then buy the connection from Billing that lights up the AI step. ## Questions people ask first Yes, and it is the reason the price reads the way it does. The receptionist answers on your own Twilio subaccount and the number you already publish, so the carrier relationship, the registrations and the caller-ID reputation stay yours, and the minutes stay on the bill you already negotiated. The call keeps moving. The caller advances to the next step you wrote down, which is usually your own voicemail, so what happens next is the thing you chose to happen next. It is worth writing that last step with some care. Receptionist is the inbound plan: it answers the phone. When you want the agent placing calls as well, attach the receptionist to an Engage, Workforce or Sales subscription for the same monthly money. The inbound connections and the outbound sessions never draw from one another. You press play on the call record. An inbound voicemail is kept there as audio, alongside the rest of that call’s evidence. $250 a month, plus $150 for each connection you added. That is the list. The calls it answers add nothing per call and nothing per minute, so a busy month and a quiet one cost you the same. Carrier minutes stay on your own Twilio bill, at the rate you negotiated. ## Build it before you buy it Write the whole answering document in a free sandbox, before a number points anywhere. Weighing it against a per-call answering service?The two models, compared. --- # What we publish, and where to find it | Vocapable > Everything Vocapable publishes so you can evaluate it without a call: the rate card, a campaign cost calculator, a sample evidence file and the agreement set. URL: https://www.vocapable.com/resources Resources # What we publish, and where to find it Judge this product on the material. The rates we will charge you, the record we write for every call, the agreements you would sign, and calls you can listen to right now. Every item is a public page or a public file, and not one of them asks for your email first. ## Price it without talking to us Both of these are on the pricing page, and neither of them ends in a form. You can build the number your finance team will ask for and leave without telling us you were here. /pricing ### The published rate card Every plan, the usage charges that sit on top of them, the volume discount and the floor it steps down to, all printed on the page. Read it, build your comparison sheet, and know the number before anyone at this company knows your name. For: Procurement, and anyone building a comparison sheet across vendors who is tired of writing “contact us” in a cell. /pricing#calculator ### Campaign cost calculator Put in your list size, how long you want to work it, attempts per contact, minutes per conversation and the answer rate you expect. It prices your campaign against the published rates, in your browser, as you type. Every default is yours to replace with your own numbers, so what comes out is an estimate you built rather than a quote we handed you. For: Whoever has to put a defensible number on a budget line before anyone will approve the spend. ## Check what we actually write down The artifact first, then the anatomy of it, then the pages that explain what produced it and what you would be signing. This is the group a compliance reviewer came for, and it is deliberately the longest one on the shelf. /media/evidence/sample-evidence-bundle.json ### Sample evidence bundle The compliance record written for a call attempt, as a JSON file you can open in an editor and hand to your own engineer. It is synthetic: a file we authored to show the shape, not an export the product generated, and no customer data of any kind is in it. For: The engineer or reviewer who would rather read the object than read about it. /evidence ### The per-call evidence file The same record rendered key for key, with the part that matters explained: what bound the dial, which checks ran, and which were waived, each waiver carrying the document it stands on and the officer who signed it. When somebody asks how a call happened, that file is the answer. For: Compliance and legal review, and the buyer who has been told “it’s all logged” by every vendor in this category and shown it by none. /compliance ### The compliance floor What runs in code on every dial, on every plan, at no charge, with nothing on our side able to switch it off: not an API field, not a plan tier, not a support ticket. Plus the gate every dial passes before it becomes a call, written as engineering with the qualifiers attached. For: The person who has to answer “what stops a bad call from going out” in front of their own leadership. /trust ### The Trust Center The security and data posture, control by control, written for the person whose job is to find the gap. Most of a security questionnaire is answered there before you send it. For: A security review, and the questionnaire that always follows it. /legal ### The agreement set Every document, in full: the master agreement, the addenda it incorporates, the waiver signed only on election, and the site policies. Each one is published from the same source text the product serves for acceptance, so the version you read here is the version you accept, hash-verified. For: Your own lawyer, before the first call rather than deep into a redline. ## Hear a call before you commit to anything The published demo calls are scripted simulations in AI voices, labelled as such on the player, and every one prints its transcript beside it: every turn, in order, so you can read what you just heard instead of taking our summary of it. Listen for the first sentence. The AI disclosure plays before the first model token on every call the product places, and the demos open the same way, because a demo that skipped it would be demonstrating a different product. Another published call, a qualification conversation from start to handoff, plays on the talk-to-sales page, and the call playing here also runs in context in the use-case section of the home page. Same label, same transcript, no email required. Play it here A cold call, start to handoff Simulated demo · AI voices - agent ▸ “Hi Dale, being upfront: I’m an AI assistant calling for Crestway Software. We help trucking companies cut dispatch paperwork. Worth thirty seconds?” - lead ▸ “Uh, go ahead.” - agent ▸ “Who runs dispatch and invoicing today? Is that all in-house?” - lead ▸ “My office manager. Takes half her week, honestly.” - agent ▸ “That’s the exact fit. Here’s my offer: we build a short demonstration around your operation, your lanes, your paperwork, and send it over. Fair?” - lead ▸ “Yeah, I’d look at that.” - agent ▸ “Done. And once you’ve had a look, when’s good to walk through it - mornings any use to you?” - lead ▸ “Tuesday after nine works.” - agent ▸ “Tuesday after nine - noted, with everything else you told me. It’s in front of the Crestway team the second we hang up, and they’ll lock the time in. Thanks, Dale.” Scripted and synthesized: AI voices reading a written conversation, so nobody on this recording is a real prospect. What it shows you is the shape of a call. ## For machines, and for the people who send them Assistants are going to be asked about this company whether or not we make it easy, so the same facts the pages carry are published in a form a model can read whole. /llms.txt ### llms.txt A structured index of every page on this site with a one-line description of each, plus the company facts and the published pricing in prose. It is generated from the same registry the pages themselves are built from, so it cannot drift away from what the pages say. /llms-full.txt ### llms-full.txt The full text of every page in one file, assembled from the built HTML after the templates, the markdown and the data files have collapsed into the bytes a visitor actually receives. https://docs.vocapable.com ### Developer docs The reference for the API, built straight from the published contract artifact rather than written alongside it, so what it documents is what the API serves. For: the engineer who has been handed the integration and wants to know what they are in for. ## Read all of it, then open an account and check the rest yourself. The sandbox costs nothing and runs the whole builder: no card, no carrier account. If a question on this page went unanswered, write to us and a person answers. ## Questions people ask about this page No. The complete rate card is published on the pricing page: every plan, the usage charges on top of them, and the volume discount with the floor it steps down to. A calculator on that same page prices your own campaign shape against those published rates, in your browser, so you can walk into your budget meeting with the number already in hand. What it produces is an estimate you built from your own assumptions rather than a quote from us. No. It is a synthetic file Vocapable authored to show the shape of the compliance record written for a call attempt. It is not an export the product generated, and no customer data of any kind is in it. It is offered so an engineer or a reviewer can read the object itself instead of reading a description of one. No. They are scripted, simulated conversations in AI voices, labelled as simulations on the player itself, and each one publishes its transcript beside it, every turn in order. They open with the same AI disclosure the product plays before the first model token, which is the part worth listening for. No. Every item here is a public page or a public file, and none of them asks for an email address, a company name or a phone number first. Open what you want, read it on your own time, and talk to us when you have a real question. Yes, and this site publishes files built for exactly that. /llms.txt is a structured index of every page with a one-line description of each, generated from the same registry the pages are built from so it cannot drift from them. /llms-full.txt carries the full text of every page in a single file, assembled from the built HTML. --- # Your Twilio, your numbers, your carrier bill | Vocapable > Your calls leave the Twilio account you already own, at the rate you already negotiated. Twilio bills you for the minutes. Vocapable bills you for the software. URL: https://www.vocapable.com/telephony Platform · Telephony ownership # Your Twilio rate is the rate you pay Every production call leaves a Twilio subaccount in your name. Twilio bills you for the minutes at the price you negotiated, straight to you. Vocapable bills you for the plan and the AI time, and that is the entire list. The numbers, the registrations and the reputation they earn stay yours, and they go where you go. ## What stays yours You already pay a carrier. Here is what each piece of that is worth the day an AI agent starts dialing on it. ### The account A Twilio subaccount in your name, on your commercial terms. The credentials you hand the platform sit in an encrypted vault, and every dial resolves them from your account. ### The numbers Bought and controlled by you. Each one connects under your account, and that account decides its credentials and its pacing budget. ### The registrations Your 10DLC brand and campaign registration, your FTC Subscription Account Numbers, your Robocall Mitigation Database entry where it applies, your STIR/SHAKEN attestation and your caller-ID authority. Your registrations are the ones a carrier looks at, so nothing here is waiting on ours. ### The rates The per-minute price you negotiated with your carrier, unchanged. It is the same price on the day an AI agent starts calling on it as it was the day before. ### The reputation Caller-ID reputation builds on numbers you keep, so it comes with you wherever the software goes. ### The authority You are the originating party on your own calls, which is another way of saying the account answers to you. Traceback and regulatory inquiries come to you, and the evidence file for every call is already in your hands. Your own contract says it in one line: you are the originating party and the party on whose behalf the calls are placed (MSA §6.1). We publish the contract, so you can read the sentence instead of taking our word for the summary. ## Twilio bills you for the minutes. We bill you for the software. Your carrier invoices you directly, at the rate you negotiated with it. Our invoice carries your plan and your AI usage. That is the whole list, and it is the reason the per-minute number below is as small as it is. The published rate is $0.13 a minute month-to-month and $0.09 on a signed commitment term, before the volume discount takes it further. It prices the AI session: the model, the speech, and the session time on our own hardware. Your carrier prices the carrier leg, which is why our rate is the same whoever carries the call. The whole rate card is on the pricing page. Every figure is published, so you can price a campaign before you talk to anybody. Where each charge comes from - Yours The minutes, the number rental and the registration pass-throughs, invoiced to you by your own carrier on your own agreement with them. - Ours The plan, and the AI usage rated against it. That is the list. Two bills from two companies, each for the work that company actually did. You can check both of them line by line. Whose account carries what ## Every production call leaves your account Our own carrier account exists for one thing: a paid test call to a destination proved through code verification. Your traffic leaves the subaccount in your name, all of it. That is why the only caller-ID reputation in this conversation is yours, and why nothing we own ever sits in the middle of your calls. The dialability gate enforces that in code, not on a policy page. Your own contract says the same, two clicks from here: MSA §6.2 and the Platform Test Calls Addendum. Vocapable runs on Twilio Media Streams. That is a fact about our stack rather than a partnership. ## Connecting is a step, not a decision You connect the account you already have, and the platform proves it works before it will dial on it. Four steps, in the portal, in this order. - 1 Connect the subaccount Add the Twilio subaccount you already have. The account identifier and auth token go straight into an encrypted vault, and every credential a dial needs resolves through your account. - 2 Verify it Verification is proof rather than a checkbox. The account reaches verified when the credentials prove out, the subaccount is reachable, and the platform stores a conservative pacing floor for the connection. You find out your account is ready before a customer does. A sandbox account carries mock evidence and stays in the sandbox, so a practice run is never mistaken for the real connection. - 3 Connect a number A number you own connects under that account and becomes a record of its own rather than a string in a config field. That account decides its credentials and its pacing budget. If the number should answer as well as call, you choose which published agent version receives it. - 4 Take the signed callback URL The platform hands you the exact signed inbound Voice URL to point the number at. When your carrier calls it, we answer only once the public callback URL, the full form body and the Twilio signature verify against your account’s vaulted auth token, and only for an active, carrier-confirmed voice number on your account with a receiving agent selected. Nobody makes your agent talk by guessing a URL. When something on your carrier account goes wrong, you see it. A suspension, a rejected registration, or a calls-per-second ceiling lower than the account claims each shows up as an error on the account and on the attempt it stopped, so you are reading a message instead of guessing at a silence. ## Workspace or headless, on the same account Use the dashboard, use the API, or use both in the same afternoon. Either way the call goes out on the account you just connected. ### Workspace Send your data as a CSV or through the API, and Vocapable becomes your calling system of record: contact records, lists, campaign boards, outcomes, dispositions and exports, with every call that ever touched a contact on one timeline. The CRM → ### Headless Your own stack stays the system of record. Contacts are created through /v1, calls are driven programmatically, and enabled outcome events can reach your configured signed webhook. The dashboard stays available as a read-only window; signup, onboarding, and first-key setup run in the browser. The API → Same engine, same /v1, same compliance floor, same pricing. The only thing that changes is where your contacts live. The trust close ## Your call, your account, your evidence Every production dial leaves your account, aimed at a list you chose. The proof of what happened lands with you as well. The calling windows, the frequency caps, the consent standard and the recording-consent class are yours to set. Set them yourself, or adopt one of the starter sets we publish. Either way we apply them on every attempt and stamp the values that bound the call into its record, along with who chose them and when, so what you hand a regulator is your own policy and the proof that it ran. The floor does not move because the account changed. The AI disclosure plays before the first model token, an opt-out writes your internal do-not-call list at once, that suppression is applied at the batch run and again at dial time, and every number is normalized. On your carrier account exactly as on ours, on every plan, and no flag anywhere turns it off. Vocapable serves US calling. The Trust Center lays out who holds what. ## What getting started actually looks like Sandbox first, your carrier in the middle, a signed order before production. The rate card is on the pricing page, and the two ways to buy a phone call are compared onBYO carrier vs resold minutes. ### Start in the sandbox A free sandbox account gets you the whole builder, and no card. It runs your script down the dial path with every gate live, before a number points anywhere. ### Connect what you already have Your Twilio subaccount, your numbers, and the registrations already sitting on them. It is the step with the most waiting in it, and it is the step that keeps the rates yours. ### Production on your carrier Going live is a signed Production Services Order and your own subaccount, connected and verified once in the portal. From then on Twilio bills you for the telephony and Vocapable bills you for the software. ## Bring the account you already have. ## Questions people ask first Yes, and it is the reason the pricing works the way it does. Every production call leaves a Twilio subaccount you own, so the numbers, the registrations, the rates you negotiated and the caller-ID reputation those numbers earn all stay yours. Connecting and verifying it once is a step in setup. Your plan and your AI usage. Twilio invoices you for the telephony directly, at the rates you negotiated, so those are the two bills you reconcile and every figure on ours is published before you sign anything. Because it prices the AI session: the model, the speech, and the session time on our own hardware. Your carrier prices the carrier leg under your own agreement with them, which is why whose account a call rides on changes nothing about what Vocapable charges for it. Paid test calls, to a destination proved through code verification, and that is the whole of it. Every production call leaves the subaccount in your name, which is why the caller-ID reputation being built here is yours. MSA §6.2 and the Platform Test Calls Addendum say the same thing in the contract you sign. You see it. A carrier suspension, a rejected registration, or a calls-per-second ceiling lower than the account claims each appears as an error on the account and on the attempt it stopped, so you are reading a message instead of guessing at a silence. No. The compliance floor is the same on your carrier account as on ours: the AI disclosure plays before the first model token, an opt-out is detected and durably written to your internal do-not-call list, that suppression is applied at the batch run and again at dial time, and every number is normalized before anything reads it. It runs on every plan, through every entry point, and no flag anywhere turns it off. Yes, and there is nothing to negotiate. The subaccount is in your name and you bought the numbers, so leaving is a matter of pointing them somewhere else. You keep the reputation they earned on the way. --- # Trust Center | Vocapable > Tenant isolation Postgres enforces, credentials the database cannot open, a compliance record on every attempt, and a carrier account that stays in your name. URL: https://www.vocapable.com/trust Trust Center # Answers your reviewer can check Your contacts, your calls and your evidence sit behind a fence Postgres enforces, and the API connects as a role that cannot lift it. Every attempt leaves a compliance record you can export. Your carrier account, your numbers and your caller-ID reputation stay in your name, so the day you leave is an export rather than a negotiation. Nothing below is a statement about how carefully we work. Each one is a mechanism with a failure mode, which is what makes it worth a reviewer’s time and worth defending to whoever sent them. The review ## Four questions decide this A security review turns on a small number of questions, and the answers that survive it are the ones somebody can go and check afterwards. Here they are in the order they arrive, each with the short answer and where the long one lives. - 1 Can another customer’s bug reach our rows? No application code has to remember to filter. Postgres decides what a query may see, and it decides before our code gets a vote. Read the long answer - 2 Where does the carrier credential we hand over live? Sealed, in a store the database layer holds no key for. The layer that can open it never opens a database connection, and the build fails if either half imports the other. Read the long answer - 3 Who at your company can open our account? A console reading through a role that owns nothing and can write one thing: a new audit row. A session that can change anything needs a second person, and that rule is a constraint on a table. Read the long answer - 4 What does the day we leave look like? Quiet. The carrier account, the numbers and the caller-ID reputation were always in your name, and your records come out as rows through an API you already use. Read the long answer Tenant isolation ## The fence is in the database, not in our code Most vendors answer this question with an assurance about their own application, and that answer is worth exactly as much as their next code review. Ours is smaller and harder to undo. Every table holding your data carries a row-level security policy, and the API connects to Postgres as a role those policies bind. A missing filter in our code is not a breach in your account, because the entitlement was settled a layer below the query. ### Every tenant table carries a policy One policy per table and the same predicate every time: this row belongs to the tenant bound to this transaction. The binding is the first statement the transaction runs and it is scoped to that transaction, so a pooled connection handed straight to the next caller carries nothing forward from the last one. ### We connect as a role the policies bind Postgres exempts a table’s owner from its own policies, which is how isolation usually turns into a comment. So there are two connection strings and they are different roles. Migrations run as the owner. The API never does: its role owns no table, holds no bypass, and carries data privileges and nothing else. ### A query with no tenant bound raises The function the policies call raises when the tenant is unset instead of quietly matching nothing. So the failure mode of a mistake in our code is an error on our side, never another customer’s contacts on yours. The silent empty result is the more dangerous outcome, and it is the one that cannot happen here. ### A misconfigured deployment serves nothing At startup the API interrogates the role it just connected as. Is row-level security on. Does this role own the table. Does it hold a bypass. Does it reach anything it should not. Any of those and it refuses to serve. The same check runs on every readiness probe, so a task that drifts is pulled out of rotation rather than left answering requests. ### Coverage is a test, not a checklist A test reads the Postgres catalog on every build and fails if a table carrying a tenant id has no policy on it. New tables cannot quietly arrive outside the fence, which is the way isolation actually erodes: not in one bad decision, but in the fortieth table nobody thought about. ### History cannot be rewritten by the application The audit table has no update policy and no delete policy for any role in the database, so a correction is a new row pointing at the row it corrects. The application tier could not rewrite what happened even by bug, which is the property that makes an old record worth reading. That is a different kind of answer from a paragraph about our engineering culture, and it is the one we would rather be judged on. It is also the answer that keeps holding while we ship: the policies, the role, the boot check and the catalog test all still apply to code nobody has written yet. Credentials ## The most sensitive thing you hand us, and where it goes You connect a carrier account. That credential can spend your money and speak in your company’s name, so it is worth knowing precisely what happens to it, and worth knowing what would have to go wrong before anyone could read one. ### Sealed before it reaches a table Your carrier credential is sealed with AES-256-GCM under a nonce drawn fresh for that record, and the key it was sealed under is named inside the sealed bytes. Rotation is adding a key and pointing new writes at it: yesterday’s records still open, and nothing has to be re-encrypted against a deadline. ### The two halves cannot reach each other The code that talks to the database handles bytes and holds no key. The code that holds the key never opens a connection. That is not a convention somebody maintains: an import rule fails the build if either half so much as imports the other, and the rule runs before a change can merge. ### Write-only, in and out A carrier credential is never returned by an endpoint, an export, a log line, or an error body. It is typed so that a request body reaching a stack trace prints as asterisks. The only readback anywhere is whether one is stored. A test compiles the actual insert statement and asserts no credential ever gets bound into it. ### Recordings under a key of their own Call audio sits in a bucket with public access blocked, transport required, and its own rotating key. A write that does not name that exact key is denied by the bucket rather than accepted quietly, so a future uploader misconfigured by us fails instead of putting one call’s audio somewhere weaker. ### The database has no door to the internet It runs in isolated subnets with no route out and no public endpoint, reachable only from the API’s own security group. There is nothing to scan, because there is nothing addressable. ### Your sandbox cannot dial a stranger Test and live are a property of the credential, not a field a client asserts. A test key routes every call to the mock provider, so no dial to the public network can originate from one and no billable usage comes out of one. Nobody ships a mistake into somebody’s evening. Access is shaped the same way. Every request is authenticated, keys carry scopes and rotate in place, and key management is treated as a privilege boundary rather than as record editing: a key cannot mint a wider key, cannot mint a live one from a test one, and cannot revoke a credential broader than itself. Passwords are stored with argon2, the browser holds an opaque session cookie and no credential, and every minting, rotation and revocation writes its audit row in the same transaction as the change. Our own people ## Who can open your account, and on whose signature This is the question a reviewer expects a vague answer to, because most vendors give one. The honest version is that support has to be able to see something, and the whole question is what that something is, who authorized it, and what it left behind. So the staff plane is built as a separate set of privileges rather than as an elevated version of yours. It reads through its own database role. It cannot write your rows. The one thing it can add to the schema is a record of itself. ### A different role, with one write in the whole schema The staff console reads through a database role that owns no table and holds no write on your data. Its single write grant anywhere in the schema is an insert into the audit log. It is not a member of the runtime role, and the API refuses to start if those two ever become the same principal. ### Staff never hold your API key Support access mints a credential of its own class instead. The secret comes back exactly once, in a URL fragment, so it never lands in a server log, a CDN access log, or a referrer header, and only its SHA-256 is stored. Nothing in that path can be replayed from a log line later. ### A session that can change anything needs two people It requires a live approval from a different staff member, and different is a check constraint on the approvals table rather than an if statement somebody could edit. The read-only session is the default, and the one that can write has to be asked for and granted. ### The reason is typed before the action, not after Every console action carries the operator’s own words for why, captured with the action rather than prompted afterwards, because a reason invented later is a caption. The change and the row that records it commit together or neither lands, so the trail cannot disagree with the database about what happened. Every call ## Controls with no off switch, and a record for each attempt Compliance here is product behavior rather than paperwork. Every line below holds on every plan, for every tenant, through every entry point, and each one is something your own team can watch happen on a sandbox account before you sign anything. ### Four controls, and nothing switches them off The AI-disclosure opener. Opt-out detection that writes your internal do-not-call list mid-call. Suppression against that list at the batch run and again at dial time. Phone normalization. Not because a policy forbids turning them off: a compliance mode can only name checks drawn from a waivable set, and these four are not members of it. Waiving the opener is not a value anyone can construct, express in JSON, or store, and two tests fail the build if that ever stops being true. ### The opener is an interlock, not a convention The model adapters live behind a gate the call driver holds no reference to. Reaching the model before the opener has played raises instead of dialing, and a call whose opener cannot play ends. So the sentence a regulator attributes to whoever originated the artificial voice is the one sentence in the system that cannot be prompt-injected away. ### Every attempt answers for itself One compliance record per attempt, written as it happens: which checks ran, which your signature covered and who signed it, the window and caps evaluated in the contact’s local time, the consent record relied on, and why the call went out or was stopped. It sits on the attempt row and comes out through the exports API with the rows intact. ### The refusals are records too A dial stopped at the gate is written with the same care as one that connected, because the call you did not place is the one you will be asked about. The gate re-runs as the channel is taken and stamps its result on the passing path as well, so a dial re-verified at launch is distinguishable from one that only ever cleared queueing. ### Uncertainty stops the dial Only an affirmative pass authorizes a call. A check that cannot complete blocks it, and a stage that did not run is never written as a stage that found nothing. Your record can tell checked, clear, and nobody looked apart, which is the difference between evidence and a log. ### Consent is a record, not a checkbox Every dial stands on a consent record: its class, the purposes it covers, where it came from, and a SHA-256 of the proof behind it, read again at dial time instead of trusted from an import. A campaign whose purpose falls outside what that person agreed to is suppressed before it dials, and the record the call stood on is named in the file. ### Stop means stop, from the same second A mid-call opt-out is a deterministic path the model does not get to argue with: the phrase list and the keypress are code constants, and the suppression commits before the call ends rather than at teardown. The next dial reads it live. It lands in a ledger with no channel column and no expiry, so it cannot fall out of step with itself. ### What you signed, hash for hash Acceptances are append-only and each one records the document version, the SHA-256 of the exact wording presented, the signer, and a server-stamped address and time. This site and the product serve one source document, so what a signer accepted and what we publish cannot drift apart afterwards. A question about a call from March becomes a lookup rather than an investigation, and the answer comes out of the record rather than out of somebody’s memory. The mechanism is written up in full on the compliance engine, and the record itself is rendered key for key, including a refused attempt, on the evidence page. Ownership and exit ## Nothing here is expensive to leave Lock-in is a procurement question before it is a technical one, and in outbound calling the expensive things to move are the carrier account, the numbers, the registrations and the reputation attached to them. All four are already in your name, so a decision to stop is a decision, not a project. ### Your carrier, your rates, your caller ID Every call a stranger answers goes out on the Twilio account you own: your numbers, your registrations, your negotiated rates, and the caller-ID reputation attached to them. Twilio bills you those minutes at the rate you already hold, so a Vocapable invoice carries no carrier line and no number rental at all. ### And no setting anywhere changes that A production dial routed onto a Vocapable-owned number is refused at the gate. It is not a mode a plan can select or a flag support can flip: the refusal sits outside the waivable vocabulary entirely, and a test walks every purpose, every consent basis and every mode to prove no combination reaches it. Your MSA §6.2 puts the same limit in the contract. ### Your records come out as rows CSV export jobs cover supported record types: contacts, contact attempts, calls, survey responses, shift confirmations, appointments, internal do-not-call entries, scrub results, usage events, and caller requests. A completed job provides a signed download for its time-bounded window. Available recordings and transcripts are fetched through authenticated /v1 call reads. ### The dashboard is optional, and always was Product resources use /v1, with cursor pagination on every list and a configured signed webhook endpoint for enabled events. Signup, onboarding, and first-key setup use browser flows. If your stack is the system of record on day one, nothing about leaving is an exercise: you were already holding the data. ### Cold outreach, on your own data, signed for Lists you purchased, rented, or had appended are yours to call. They run under the Cold Outreach & Data Provenance Waiver you execute, which records that the data is yours under your own supplier agreement and that the consent basis, the jurisdictions and the responsibility for that program stay on your side. Scraped and unauthorized-extraction contacts are refused however they arrive, by upload or by API. Every import carries a signed attestation of where the contacts came from, the four controls above are unchanged by the waiver, and every attempt writes its record either way. ### Cards never touch us Payment is handled by Stripe. The browser never receives a secret from us, and no card detail is stored on our side, so the part of your review that is about payment data is a short conversation. The same shape holds on the way in. Bring the account you already have, keep the rates you already negotiated, and run the whole dial path against simulated telephony first, every gate live and failing closed, before a single number rings anywhere. See how the telephony works and what the plans cost. The division ## Clear lines beat warm assurances Here is the division, so nothing in your posture rests on an assumption about ours. A signature records each half, which is why an answer six months from now can name a document and a person rather than a recollection. ### One signature, riding every record Before the first production dial you sign an attestation naming the screening arrangement you already run. It does not sit in a folder after that: its reference and its signer ride into every record, so “who screened this number, and on which document” is answered on the call itself instead of in somebody’s email. Keep the supplier you have and the pricing you negotiated for it, and never pay twice for a lookup you already buy. ### What happens after the call is yours to run Each event you enable is signed and retried. A call.ended event arrives at call completion with how the call ended, its duration, related record identifiers, and evidence availability; call.analyzed adds post-call analysis later. Survey and shift outcomes arrive through their own signed events and resources. Your rep or coordinator can fetch available call evidence through authenticated /v1 reads and work the next step in the Inbox. ### The law where you dial is your call to make The platform enforces the policy state on every attempt, quiet hours and frequency caps and consent standards, and records the values it evaluated, so the proof of what was enforced is yours to produce on demand. You choose who gets called, and the duty that choice carries where you dial stays with you: what the law requires there is determined by you and your advisors, and we say so here rather than in a footnote. ### Your data, and it leaves whenever you want it to Call recordings, transcripts, and compliance records are processed on your behalf as a service provider. Retention, deletion, and what happens to the audio when an agreement ends are written into the Data Processing Addendum and the Privacy Policy, and both are published in full below before anybody signs. ## Scope of service Where the service begins and ends, stated plainly so your reviewer does not have to infer it from what a page happens to mention. - Vocapable serves US calling today, and the privacy machinery we operate is built on CCPA and CPRA. - Calls involving protected health information are outside our Acceptable Use Policy. - Campaigns run on the contacts you supply, captured or acquired, and every import carries a signed attestation of where they came from. - When caller voicemail is recorded, it is available on the call record for your team to play. Running a security review ## Send the questionnaire. You get a written answer, from the person who built it. Mail it to support@capstralabs.com and you get direct answers about the controls themselves rather than a portal invitation. We publish the control descriptions instead of holding them back for a questionnaire, which is why most of your list is already answered above and the rest is a reply rather than a process. The shortest version of the whole thing, written for the person who has to summarize it upward, is here: how your customer data is protected. ## Questions reviewers ask first Postgres, before our code runs. Every table holding tenant data carries a row-level security policy keyed to the tenant bound to the current transaction, and the API connects as a role those policies bind: it owns no table, holds no bypass, and has data privileges and nothing else. A query that reaches those tables with no tenant bound raises rather than returning rows, so the failure mode of a bug on our side is an error on our side. A test reads the database catalog on every build and fails if a table carrying a tenant id has no policy on it. The API refuses to serve. At startup it interrogates the database role it just connected as and stops if row-level security is off, if the role owns the table, if it holds a bypass, or if it reaches anything it should not. The same check runs on every readiness probe, so a task that drifts after startup is pulled out of rotation instead of left answering requests. Failing at boot is the behavior you want from an isolation boundary, because the alternative is a deployment that looks healthy and is not. Sealed with AES-256-GCM in a store whose database half holds bytes and no key, and whose key-holding half never opens a database connection. An import rule fails the build if either half imports the other, so the separation survives the next refactor rather than depending on somebody remembering it. The credential is never returned by an endpoint, an export, a log line, or an error body, and a support view of your account carries the address of the record rather than anything inside it. The staff console reads through a database role that owns no table and whose only write grant in the whole schema is an insert into the audit log, and the API refuses to start if that role and the runtime role are ever the same principal. Staff hold no customer API key: support access mints a credential of its own class, returned once and stored only as a hash. A session that can change anything requires a live approval from a different staff member, enforced as a constraint on the approvals table, and every action carries the operator’s own stated reason, written in the same transaction as the change it describes. No, and not because a policy forbids it. A tenant’s compliance mode can only name checks drawn from a waivable set, and the four floor controls are not members of it; the unwaivable stages are derived by set difference rather than listed, so they cannot be omitted by an edit. Waiving the disclosure opener is not a state anyone can construct, express in JSON, or store, and two tests fail the build if that stops being true. They are also never metered: the local stages are excluded from billable lookups by construction, so there is no configuration in which the floor is charged for or switched off to save money. An export, on your own schedule. The Twilio account, the numbers, the registrations and the caller-ID reputation were always in your name, so nothing has to be migrated off ours. CSV jobs cover supported record types: contacts, contact attempts, calls, survey responses, shift confirmations, appointments, internal do-not-call entries, scrub results, usage events, and caller requests. Available recordings and transcripts are fetched through authenticated /v1 call reads. Product resources use /v1, while signup, onboarding, and first-key setup remain browser flows, so a team that integrated on day one is already holding its own records. Written answers about the controls themselves, from the people who built them, at support@capstralabs.com. The control descriptions are published rather than held back for a questionnaire, which is what this page is: most of what a reviewer asks is already answered above, and the rest is a reply rather than a process. ## Read the whole set before you sign any of it Every document that forms the agreement is published here in full, and the product serves the same source text, hash for hash. Nothing is incorporated by reference into a page you cannot open, and nothing arrives for the first time in a signature request. - Master Services Agreement - Telecom Compliance Addendum - Platform Test Calls Addendum - Acceptable Use Policy - AI Technology Disclosure - Data Processing Addendum - Cold Outreach & Data Provenance Waiver - Privacy Policy - Website Terms of Use ## Check it on your own account Open a sandbox, run the whole dial path against simulated telephony, and read a compliance record from a call you ran yourself. The floor runs there exactly as it runs in production, and nothing is pointed at anyone. --- # CSAT & feedback calls | Vocapable > Feedback calls in your voice: collect structured survey responses from completed feedback calls and use CSAT and NPS views to follow customer experience. URL: https://www.vocapable.com/use-cases/csat-surveys Use case · CX & customer voice # Hear how your customers really feel, on a call, on the record. You call the people you already serve, from the number they recognize, about a transaction that actually happened. The agent asks your questions in your voice and returns structured survey responses from completed feedback calls, while the interaction is still useful to your team. A customer at home answers a call with a genuine smile (Illustrative footage.) ## From the day’s transactions to a call sheet your team can act on - 1 Your list, your consent basis You call the people you already serve about a transaction that happened, so the basis you declare is one your own records can back up. It rides on the contact under your signed import attestation, ready to produce. - 2 The conversation, your questions The agent opens by disclosing it’s AI, asks your questions in your voice, and listens. Two questions, thirty seconds, the day the visit happened, not a form three weeks later. - 3 A structured response on the record For completed survey responses, the completion state and answers are saved with the call and contact. Scored questions add a score, so your team can work from customer feedback instead of a stack of callbacks. - 4 Delivered where you work For enabled events, call completion can reach your configured signed webhook, while supported CSV exports and /v1 reads give your team survey responses linked to the call and contact. The CRM → The API → - 5 Your team acts on it Your team can route a follow-up from the result while the interaction is still fresh. The call sheet is yours, built from responses instead of guesses. ## It asks like you would, because it’s built from your business Your questions, in your order, with your follow-ups. Persona and guardrails keep the conversation inside what you would say yourself, and knowledge packs help it answer the practical details your team provides. Then you prove it: the sandbox runs your real script down the whole dial path with every compliance gate live and failing closed and every disposition written the way production writes it, before a number points at a single customer. What the agent knows, because you gave it - Your feedback questions and follow-ups, in order - Your persona, your tone, your voice - Guardrails that limit what the conversation may say - When to stop, when to thank, and when to flag for a human The outcome panel ## The numbers you can put in front of your team Each figure comes off the calls you actually placed, in your own account. Contact rate How many of the customers you chose to ask actually picked up and talked. Completed conversations Feedback calls that reached the end of your questions, counted per campaign. CSAT & NPS views Use scored survey responses to see customer feedback by campaign and question. Cost per completed call Read off the usage ledger for the exact calls this campaign made. For enabled events, call completion can reach your configured signed webhook. Supported CSV export jobs and a /v1 read keep survey responses linked to their call and contact. The trust close ## Your customers, your list, and a record of every call. These are people you already serve, dialed from your own carrier account. The reason you gave for calling them travels with the contact into every attempt. Every attempt passes the same checks, and they fail closed: phone normalization, your own internal do-not-call list applied at the batch run and re-checked again at dial time, and the calling window and caps read in the customer’s own local clock. Every call opens with the AI disclosure. Every call, placed or held back, writes an exportable evidence file. Those checks are on every plan at $0, and nothing in a plan or a support tool switches them off. Where the list came from is signed for at every import: the CSV import signs for it and so does the standing attestation behind every contact your API creates. A purchased, rented, or appended list is admitted under the Cold Outreach & Data Provenance Waiver you execute, and the floor is the same for it: the AI disclosure opens the call, your own suppression list applies at the batch run and again at dial time, and the evidence file is written either way. That is the file a reviewer asks for: who signed for the list, which checks ran on each dial, what the agent asked and what your customer answered. It is built call by call, so it exists before you need it. Every control behind it, one at a time, is on the Trust Center. ## What getting started actually looks like The sandbox is free and it holds the whole builder. When you are ready, choose a plan, accept the production terms, add a payment method, connect your verified Twilio account, and activate production in the portal. The rate card and the $0 compliance floor are laid out on thepricing page. ### Start with the builder A free sandbox account gives you the whole builder: your feedback questions, your persona and guardrails, your voice. Draft the conversation in an afternoon and prove it before anyone’s phone can ring. ### You pick who gets asked The list is yours to choose. Vocapable records the consent basis you declare with the contact and applies it on every attempt, so every call carries its own reason for being made. ### The number they recognize Production calls leave on your own Twilio subaccount, from the numbers your customers already know, on the caller-ID reputation you built. Twilio bills those minutes straight to you at your rate, and our rate card prices the AI session rather than the carrier leg. ## Your customers are one call away. Ask them. --- # Staffing calls at roster scale | Vocapable > Roster-scale staffing calls from your own numbers, with completed responses captured as shift confirmations and decline reasons. URL: https://www.vocapable.com/use-cases/fill-shifts-automatically Use case · Staffing & workforce # Every open shift gets a calling program your team can act on. Vocapable helps you run a roster-scale calling program from your agency’s own number. Confirmations and declines from completed conversations become structured shift outcomes, so your coordinator can mark tonight’s roster from answers. A small staffing agency office in the evening, two colleagues in easy conversation (Illustrative footage.) ## From an open shift to a marked roster - 1 Tonight’s shifts, your list The open shifts and the candidates who could work them, imported as a CSV or pushed by API, each carrying the consent basis you declared and checked against your suppression list before the first dial. - 2 A roster-scale calling program Run the outreach from your agency’s own number, with each conversation opening by saying it’s AI and following the staffing guidance your team provides. - 3 A shift outcome your team can use Each call can create a shift confirmation tied to the shift and candidate: confirmed, declined, or another response, with the reason your coordinator needs to make the next move. - 4 Delivered where you work For enabled events, call completion can reach your configured signed webhook, while supported CSV exports and /v1 reads give your team the call record and structured shift outcomes. The CRM → - 5 Your coordinator marks the roster The roster stays your coordinator’s: they act from confirmed shifts and documented decline reasons instead of working back through a voicemail queue. ## It sounds like your agency, because it’s built from it Your agency’s name and tone, your sites and their details, your escalation language for the questions that need a human. Knowledge packs turn what your coordinators already know into what the agent actually says on the call. Then you prove it: the sandbox runs your real script down the whole dial path with every compliance gate live and failing closed and every disposition written the way production writes it, before a number points at a single candidate. What the agent knows, because you gave it - Your staffing language, tone, and escalation guidance - The sites, the start times, and who to ask for at the door - Guardrails that limit what the conversation may say - When to stop, when to retry, and when to hand off to a human The outcome panel ## The numbers behind tonight’s roster Each one is counted off your own calls, in your own account. Contact rate How many of the candidates you chose to call actually picked up and talked. Shift confirmations Confirmed and declined responses, including the decline reason, linked to the shift and call. Cost per completed call Read off the usage ledger for the exact calls this campaign made. Enabled shift-confirmation events are available through a configured signed webhook. Supported CSV exports and a/v1 read include documented call details and decline reasons. The trust close ## You decide who gets called. Every call comes back documented. These are your people and your list, dialed from your own carrier account. Every dial passes the same checks, and they fail closed: phone normalization, your own internal do-not-call list applied at the batch run and re-checked at dial time, and the calling window and caps read in the candidate’s own local clock. Every call opens with the AI disclosure before the first model token. An opt-out is honored across your whole tenant the moment it lands. Every call, placed or held back, writes an exportable evidence file naming the checks and the values behind them. That floor costs $0 on every plan, and nothing in the portal or the API loosens it. It runs the same way on the night a new coordinator publishes their first campaign as it does on the night your best one does. Where the roster came from is signed for at every import. Your own crew and your own applicants are ready to call as they are; a candidate list you purchased, rented, or had appended is called under the Cold Outreach & Data Provenance Waiver you execute, which keeps the supplier agreement and the consent basis in your name. That is the file a reviewer asks for: who signed for the roster, which checks ran on each dial, what the agent said and what the candidate said back. It writes itself while the calls run. Every control behind it, one at a time, is on the Trust Center. ## What getting started actually looks like Build the whole confirmation call in the sandbox today. When you are ready, choose a plan, accept the production terms, add a payment method, connect your verified Twilio account, and activate production in the portal. The rate card and the $0 compliance floor are laid out on thepricing page. ### Start with the builder A free sandbox account gives you the whole builder: your agency’s name and tone, your sites, your escalation language, your voice. Draft the confirmation conversation in an afternoon and prove it before anyone’s phone can ring. ### The staffing call stays yours You pick the candidate list and your coordinator keeps the roster. The agent does the dialing and the documenting, so the decision gets made on answers instead of on who happened to call back. ### Go live on the account you own When you are ready, choose a plan, accept the production terms, add a payment method, and connect your verified Twilio account in the portal. Your account, numbers, registrations, negotiated rates, and caller-ID reputation stay in your name. ## Tonight’s shifts are one campaign away. Start the calls. --- # Qualify leads at scale | Vocapable > Run outreach campaigns that turn completed conversations into qualified outcomes, available in your workspace and delivered to your systems. URL: https://www.vocapable.com/use-cases/qualify-and-book Use case · Speed-to-lead # Turn campaign conversations into qualified outcomes. Upload the list, set the campaign in motion, and qualify completed conversations against your own rubric. Capture the interest, reasons, and next-step details people share, then deliver structured outcomes to the workspace and systems your team uses. A professional smiles at her desk as morning campaign results come in (Illustrative footage.) ## From your list to qualified outcomes your team can use - 1 Bring the list you own Your leads, imported as a CSV or pushed straight from your own stack through the API, checked against your suppression list before the first dial. - 2 Run the campaign on your schedule Launch, pause, resume, or cancel from the workspace. Eligible calls follow the calling window and frequency caps, and connected conversations begin with an AI disclosure. - 3 Qualified against your rubric Your qualifying questions, your disqualifiers, your definition of interested, asked in your order and written down as the lead answers them. - 4 Prove it before it dials a stranger Shape the conversation in the builder from your product knowledge and rubric, then run the whole dial path in the sandbox with every gate live before any prospect’s phone can ring. - 5 Turn completed calls into usable records Completed conversations produce a qualification, disposition, and any details the lead shares. The resulting contact record is available in the workspace and through configured signed events, supported CSV exports, or the /v1 API. The CRM → The API → ## It sells like your best rep, because it’s built from your business Start from a working draft, then make it yours: your product knowledge, your pricing posture, your qualifying rubric, your handling of “send me an email.” Upload your pricing sheet, your FAQ, your product PDFs, and a knowledge pack turns them into what the agent actually says on the call. Then you prove it: the sandbox runs your real script down the whole dial path with every compliance gate live and failing closed and every disposition written the way production writes it, before a number points at a single prospect. What the agent knows, because you gave it - Your qualifying questions and hard disqualifiers, in order - What your product actually does, in your words - Your handling of “send me an email” - When to stop, when to retry, and when to hand off to a human The outcome panel ## The numbers you can put in front of your boss Every figure here is computed on the calls you placed, in your own account. Qualification rate How many completed conversations were qualified against your own rubric, per campaign. Disposition mix Every attempt written from a pinned vocabulary, so this month reads like last month. Time-captured leads Qualified outcomes that include a time window or next-step detail the person shared. Refusals, by reason Every dial the platform held back, with the value behind it, ready the day someone asks. Conversation draft The reviewed conversation itself: your rubric, your knowledge, your voice, versioned. Call-lifecycle records Completed call records with disposition and available transcript, recording, or voicemail evidence for review and export. The trust close ## The vendor your legal review can actually read Before a dial, Vocapable applies phone normalization, your internal do-not-call list at batch run and dial time, plus the calling window and caps in the lead’s local clock. Calls begin with an AI disclosure, and opt-outs are suppressed across the tenant. Call activity, including held-back attempts, writes exportable evidence. That floor costs $0 on every plan, and no plan, portal setting, or support tool loosens it. You decide who gets called and on what basis. Vocapable records that basis with the contact and applies it on every attempt, so the answer to “why was this number dialed” is on the record from the first dial. That is the whole file a reviewer asks for: who signed for the list, which checks ran on each dial, what the agent said, and what the contact said back. It is written as the calls happen, so it is ready long before anyone asks. Every control behind it, one at a time, is on the Trust Center. ## What getting started actually looks like Open the sandbox today and build the whole conversation. You sign an order when you are ready to dial, and payment starts at production. The rate card and the $0 compliance floor are laid out on thepricing page. ### Start in the sandbox A free sandbox account gets you the full builder, no credit card. Set up your agent, your rubric, your knowledge packs, and hear the simulated demo calls. ### Bring the list you already have The motion runs on the leads you captured, and on the lists you purchased, rented, or had appended once you sign the Cold Outreach & Data Provenance Waiver. Every import is signed for where the contacts came from, so provenance is in the file from the first dial. ### Go live on the account you own A signed production order connects your own Twilio subaccount, and the calls go out on what you already built: your numbers, your registrations, your negotiated rates, your caller-ID reputation. Twilio bills those minutes straight to you at the rate you negotiated, and our rate card prices the AI session rather than the carrier leg. That is why the number is this small. ## Put your lead list to work with a campaign your team can review. --- # Voice agents & campaigns | Vocapable > Build and release AI voice agents from ready-made templates, run campaigns from your own Twilio account, and turn calls into structured work for your team. URL: https://www.vocapable.com/voice-agents Product · Voice agents & campaigns # Build the conversation once. Put every call to work. Start from a ready-made workflow, make the agent sound like your business, and run a controlled program on the Twilio account and numbers your company already owns. Turn completed conversations into useful next steps for your team. One platform, several jobs ## The call becomes part of the work, not a separate project ### Qualify demand Ask the questions that separate a real opportunity from a casual inquiry, then send the result to the team that follows up. ### Run workforce calls Check staffing availability, send shift reminders, and collect confirmations or decline reasons without asking coordinators to work a phone tree. ### Listen to customers Collect feedback in your voice, keep structured survey responses with the call, and use CSAT or NPS views to see what needs attention. ### Keep callers moving Use the same platform for an AI receptionist that answers on your own number and turns messages, tickets, and callbacks into team work. Agent lifecycle ## A clear path from your first draft to a live program The workspace gives operations teams a practical way to build, check, and run voice work without losing the history of what changed. - 01 Start with a working template Choose a starting point for staffing availability, shift reminders, candidate follow-up, CSAT, or sales outreach. It gives your team a useful structure from the first draft. - 02 Make the conversation yours Set the objectives, questions, voice, and business knowledge the agent uses. The agent speaks from the material and guardrails your team provides. - 03 Release with evidence Test a draft, then publish an immutable version with golden-transcript release checks before a changed agent goes into a program. - 04 Operate the campaign Import a list with row-level issue reports, set the program up, and control its launch, pause, resume, or cancellation from one workspace. What the team receives ## Conversations that leave your team with a next action ### Structured outcomes Lead status, shift confirmations, and survey responses are durable records your team can filter, report on, export, or receive through configured signed events. ### A useful call record Open the outcome, disposition history, and available call evidence together when your team needs context for the next action. ### An Inbox for caller requests Messages, support tickets, callbacks, and voicemail receipts appear in one operational feed for the people who need to act. ## Your next call program can start with the workflow your team already knows. Build it in the free sandbox, then choose a plan and activate production from the portal when your workspace is ready.