Cartesia Sonic TTS API + MCP by Cartesia

# Fix list: Cartesia Sonic TTS API + MCP

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/cartesia-tts, the October 2026 research run, assessed 1 October 2026. Grade B, 64.2 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 Cartesia Sonic TTS API + MCP: 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, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 (0). Plan prices and credit allowances are public, about 1 credit a character, but there's no plain dollar price per character outside the plans (10). Free plan with 20,000 credits a month (about 27 minutes), no card step mentioned, non-commercial use only (20). A person signs up in a browser (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. Reliability, 63 out of 100, up to 7.4 more on the total

Why it scored 63: incident.io status page at status.cartesia.ai with TTS components for the US, EU, APAC and Australia and a dated history (20). Five TTS incidents in the last 90 days. A partial TTS outage in the US for 41 minutes on 29 July (the postmortem counts 50 minutes of failed requests), elevated errors across the whole API for 58 minutes on 1 August, intermittent TTS timeouts in three regions from 21:25 to 23:21 on 5 August, and smaller ones on 13, 14 and 21 August. We count 5 August as one major and the rest as partial (10). TTS concurrency published per plan, 2 on Free up to 15 on Scale (15). A 429 is documented when the limit is hit, with no Retry-After or backoff guidance in the docs, though the Python SDK retries 429 and 5xx with backoff (8). No SLA found, and the Terms disclaim availability (0). `sonic-3.6` is GA, while emotion tags and `sonic-preview` are beta (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.

## 3. Security & auth, 58 out of 100, up to 7.4 more on the total

Why it scored 58: Model reading of the checklist, with training and retention in place of least-privilege and injection lines. Revocable API keys and short-lived access tokens with TTS, STT and agent grants and a set lifetime, but no key scoped to TTS. Browsers pass the short-lived token as a WebSocket query parameter, which we don't deduct for (25 of 30). The Terms (14 June 2024) let Cartesia train on inputs and outputs unless you file an opt-out form (10 of 20). Zero data retention is Enterprise only and no retention period is published for other plans (5 of 15). Usage totals in the dashboard, no per-call log found (8 of 15). SOC 2 Type II, HIPAA and PCI claimed by Cartesia and a trust centre at trust.cartesia.ai, but 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.

## 4. Agent ergonomics, 65 out of 100, up to 5.7 more on the total

Why it scored 65: API reading of the checklist. WAV, MP3 or raw PCM from 8 to 48 kHz, over bytes, SSE or a WebSocket that takes streamed LLM text, with word timestamps on request (22 of 25). Voice lists paginate with SDK iterators, filters not confirmed, and no per-request length cap is documented (10 of 20). Errors are HTTP statuses mapped to SDK exceptions without TTS-specific codes (10 of 20). The SDK retries 429 and 5xx with backoff. We found no seed and didn't confirm the billing rule for failed calls (10 of 20). Four required fields plus the version header, and official Python and TypeScript SDKs (13 of 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.

## 5. Transparency & trust, 71 out of 100, up to 2.5 more on the total

Made of editorial 60, provenance 82.

Why it scored 71: Closed service with clear terms, SDKs under Apache-2.0 (15). Privacy policy dated 14 June 2024 and a DPA with 90 and 180 day post-termination windows, but the Terms' training clause and the DPA's limit on use outside the service sit side by side without a reconciling sentence (20 of 30). Dated deprecation notices in the changelog, no written notice policy (15 of 20). Sub-processors sit behind the trust centre, which we couldn't read without a browser. Regional status components for the US, EU, APAC and Australia (10 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: cartesia.ai, registered 2023-05-10 (3 years) (7 of 15)
- security.txt: not found (0 of 10)

## 6. Schema & documentation, 86 out of 100, up to 2.3 more on the total

Why it scored 86: OpenAPI files for the latest and each dated API version, plus AsyncAPI files for the WebSocket, are listed in llms.txt (25). llms.txt and Markdown pages (10). A page compares the bytes, SSE and WebSocket endpoints and says which to use for what, and model pages explain dated snapshots (15 of 20). Typed request bodies with enums for container, encoding and sample rate, and a 0.6 to 1.5 range on `speed` (13 of 15). Examples in curl and both SDKs, but no error responses documented for the TTS endpoints, only SDK exception classes (8 of 15). Dated `Cartesia-Version` header, dated model snapshots that never change, and a monthly 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. Maintenance & community, 79 out of 100, up to 1.8 more on the total

Why it scored 79: Read as a model. `cartesia-mcp` 0.26.0 on 2026-09-30, Python SDK 4.2.0 on 2026-09-02 and the `sonic-3.6-2026-08-27` snapshot (30). Eight Python SDK releases between 17 July and 2 September (10). Sonic-2, Sonic Turbo and `sonic-3-2025-10-27` sunset after 2026-10-20, announced about 83 days ahead (5 of 10). Public changelog, no support channel confirmed (5 of 15). The Python SDK has one open issue, from April 2025, with no reply visible (7 of 10). Current Python and TypeScript SDKs (15). The Python SDK shipped within 30 days, with migration notes for the 2026-08-14 API version in 4.0.0. CI not checked (7 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.

## 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.

- We didn't open the OpenAPI file itself, only its entry in llms.txt, so the listing's `openapi` stays null.
- Whether failed requests are free. The listing says credits are charged only on success, which we didn't see today.
- Whether a card is needed for the Free plan. The pricing page doesn't mention one.
- What the trust centre lists for sub-processors and data locations.
- The listing's `toolCount` of 15 looks stale. The `cartesia-mcp` 0.26.0 page lists 18 tools by our reading, including destructive ones such as `delete_voice`, while the voice-cloning check counted 19. We left the field alone.

## Weaknesses

- Five TTS incidents on the status page between 29 July and 21 August 2026
- TTS concurrency is 2 on Free, 3 on Pro, 5 on Startup and 15 on Scale
- The Terms allow training on inputs and outputs unless you opt out by form
- Zero data retention is Enterprise only
- No SLA, and no error responses documented for the TTS endpoints

## 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.

- Send the `Cartesia-Version` header on every request.
- Pin a dated snapshot such as `sonic-3.6-2026-08-27` if output must not change.
- Move off `sonic-2`, `sonic-turbo` and `sonic-3-2025-10-27` before 2026-10-20.
- Expect a 429 at the plan's concurrency limit and queue requests yourself.
- Mint a short-lived access token for browser clients instead of shipping the API key.

## What the review panel asked for

- Publish a dollar price per character
- State billing for failed requests
- Document TTS error responses and Retry-After
- Publish an SLA

## 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.