Resemble AI Voice Cloning API by Resemble AI

# Fix list: Resemble AI Voice Cloning API

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/resemble-ai-voice-cloning, the October 2026 research run, assessed 1 October 2026. Grade C, 54.3 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Resemble AI Voice Cloning API: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Payments & pricing, 20 out of 100, up to 10 more on the total

Why it scored 20: No x402, MPP or L402 (0). Per-unit prices are published without a login, $0.0005 a second of speech on Business and $1.50 a month per clone slot (20). No API cloning below Business, and the free first clone in the Flex web app captures a card for later clones (0). Signup is a human browser flow (0).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):

The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).

- 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which.
- 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login.
- 20, a free tier or trial that doesn't need a card.
- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).

Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied.

Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol.

## 2. Security & auth, 54 out of 100, up to 8.1 more on the total

Why it scored 54: Graded for voice cloning, with consent and misuse controls in place of the read-only line and training and retention of voice data in place of the prompt-injection line, since the API returns audio and IDs rather than third-party text. Plain Bearer API keys, no scopes found (20 of 30). The terms say Resemble may require verbal consent from the person cloned. The API reference has no consent field, though the Node SDK still sends an optional `consent` value added in 2023. Watermarking and deepfake detection sit in the same account (4 of 20). The privacy policy of 14 August 2026 says customer voice data isn't used to train general-purpose models or for research, recordings and voice models are kept while the account is active plus a 30-day grace period, and synthesis text and audio aren't kept after delivery (12 of 15). Clone slots and usage readable through the Billing API, no per-call log found (8 of 15). The Secureframe trust centre lists ISO 27001:2022 as certified, an annual third-party penetration test, and SOC 2 Type 2 still in its observation period on 2 October, two days after the date it gave for the report. No security.txt, disclosure policy or bug bounty found (10 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-security):

- 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option.
- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.
- 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10.
- 0 to 15, audit logs or per-call visibility for the operator.
- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.

Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing.

## 3. Reliability, 65 out of 100, up to 7 more on the total

Why it scored 65: Status page at status.resemble.ai (Checkly) with component history (20). The current components cover the web app, Resemble Ultra TTS and the safety APIs. The history shows a 'Chatterbox Turbo Rapid Voice Clones' component in July that's no longer listed. Since early July the voice surfaces had only short incidents, several of 1 to 5 minutes on rapid voice clones from 6 to 9 July and 1 minute on Ultra TTS on 22 September. The Identities API, a separate product, has had an incident open since 9 September, and Detect had a 50-minute incident on 27 August. Minor incidents only on the voice surfaces (20). 40 requests a second per API token on the rate-limits page, 10 a minute on audio enhancement (15). The rate-limits and error pages say nothing about 429, Retry-After or backoff (0). No SLA found (0). Rapid and professional clones are generally available (10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):

Hosted APIs, MCP servers, models and platforms.

- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).
- 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so.
- 15, rate limits documented with numbers.
- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.
- 10, an SLA published for any paid tier.
- 10, the surface agents use is generally available, not beta or preview.

Local packages, SDKs, frameworks and stdio MCP servers.

- 20, installs from an official package with supported runtimes stated.
- 25, a public CI and test suite, passing on the default branch.
- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).
- 15, semver discipline and breaking changes called out in a changelog.
- 15, version 1.0 or later, or declared stable.

Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.

## 4. Maintenance & community, 30 out of 100, up to 6.1 more on the total

Why it scored 30: Newest changelog entry 2026-06-30, 94 days before this recheck (10 of 30). No dated entries in the last 90 days (0). Public changelog, no answering support channel confirmed (5 of 15). Both official SDKs were last updated on 6 April 2026, Node 3.5.3 by commit and Python 1.9.0 on PyPI, after Python 1.8.0 in February. Neither has a release since the June model changes (12 of 15). The Node repo has Jest tests but no CI workflow and no release tags (3 of 10).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):

- 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older.
- 20, at least three releases or dated changelog entries in the last 90 days.
- 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15.
- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).
- 10, package health, current dependencies and CI.

Models are read for deprecation notice periods and model churn rather than release counts.

## 5. Agent ergonomics, 66 out of 100, up to 5.5 more on the total

Why it scored 66: Compact voice objects (20 of 25). Voice lists page with `page` and `page_size` up to 1,000, and `sample_url` and `filters` flags are optional (15 of 20). Errors come back as `success: false` with a message and no documented codes, with advice to log the request ID for support (6 of 20). No idempotency key, but builds report through a `callback_uri` webhook so nothing needs re-posting (10 of 20). Official Node and Python SDKs and few required fields (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):

- 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries).
- 20, pagination, filtering and output-size controls.
- 20, actionable, documented error responses, codes and messages an agent can recover from.
- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.
- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.

Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs.

## 6. Schema & documentation, 81 out of 100, up to 3.1 more on the total

Why it scored 81: Public OpenAPI 3.1 at docs.resemble.ai/openapi.json (25). llms.txt (10). The clone overview separates rapid and professional clones and says how much audio each needs (12 of 20). Typed create fields, including `dataset_url`, `callback_uri` and `language`, and `page_size` bounded to 10 to 1,000 (12 of 15). Examples, plus dataset quality scores (STOI, PESQ, SI-SDR) for bad recordings. The spec documents only 200 responses on the voice, build and delete operations, and the errors page gives one generic shape, `success: false` with a message (7 of 15). Versioned `/api/v2` paths and a dated public changelog (15).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):

APIs and MCP servers.

- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).
- 10, llms.txt or Markdown docs served for agents.
- 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference.
- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.
- 0 to 15, examples and documented error responses.
- 15, versioning and a public changelog.

Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference.

## 7. Transparency & trust, 78 out of 100, up to 1.9 more on the total

Made of editorial 70, provenance 86.

Why it scored 78: Closed service with clear terms (15 of 30). The privacy policy of 14 August 2026 states no training on customer voice data, retention while the account is active plus 30 days, no retention of synthesis text or audio, and points to a DPA and a named subprocessor list. That agrees with the terms, which have Resemble destroy voice models on request after termination (25 of 30). The changelog's 29 June 2026 entry says Resemble 'began deprecating all older voice models', with no notice period. The model versions page now lists seven end-of-life TTS models, Chatterbox and Chatterbox-Turbo among them, and says voices still on them can no longer generate audio until upgraded to Resemble Ultra, but gives no dates or notice policy (10 of 20). Subprocessors named at trust.resemble.ai/subprocessors, and hosting stated as the United States on Google Cloud us-east4, Render in Oregon and Cerebrium (20 of 20).

The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):

- 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms.
- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).
- 0 to 20, a deprecation policy or notices with dates.
- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).

The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.

Provenance checks not met in full (half of this category, computed from checked facts):

- Domain age: resemble.ai, registered 2018-11-12 (7 years) (11 of 15)
- security.txt: not found (0 of 10)

## Deductions

Each comes off the total. A fixed and documented problem counts for less at the next check.

- 2026-06-29: the changelog says Resemble 'began deprecating all older voice models' that day, with no model list or notice period, and the model versions page now says voices on the seven end-of-life models can no longer generate audio until upgraded to Resemble Ultra. We found no earlier public notice. An upgrade path exists, and we can't see whether customers were told privately, so a small deduction. https://www.resemble.ai/changelog and https://docs.resemble.ai/getting-started/model-versions

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- When the seven end-of-life TTS models stopped generating audio, and whether customers had notice before the 2026-06-29 changelog entry. The model versions page gives no dates
- Whether the SOC 2 Type 2 report, expected by 30 September 2026 per the trust centre, exists. The trust centre still showed the observation period on 2 October
- Whether the optional `consent` value the Node SDK sends is still accepted or checked. The create-voice reference lists only name, dataset_url, callback_uri and language
- Why the 'Chatterbox Turbo Rapid Voice Clones' status component from July no longer appears among the current components. Chatterbox-Turbo is among the end-of-life models, which may explain it

## Weaknesses

- Cloning API only on Business ($1,000 a month) and Enterprise
- No consent or speaker verification step in the API reference
- Seven older TTS models are end of life and their voices can't generate audio until upgraded. The 29 June 2026 changelog entry gave no notice period
- Error responses carry no codes, create-voice documents only a 200, and nothing is documented for 429
- No security.txt, disclosure policy or bug bounty, and SOC 2 Type 2 was still in observation two days after its expected report date

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Create the voice, upload recordings or pass `dataset_url`, then call `/build`. Nothing trains until you do
- Set `callback_uri` rather than polling for `status` to reach `finished`
- Expect a rapid clone. Since 2026-06-30 API-created voices default to rapid, and create-voice has no field to pick another type
- Upgrade any voice on a pre-Ultra model before synthesis, since those voices can no longer generate audio
- Check `success` in every response body, log the request ID, and stay under 40 requests a second per token with your own backoff

## What the review panel asked for

- Voice design endpoint
- Cloning on cheaper plans
- enforced consent field
- scoped keys

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.