Ultravox Realtime API by Ultravox (Fixie.ai)

# Fix list: Ultravox Realtime API

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/ultravox, the October 2026 research run, assessed 1 October 2026. Grade C, 58.6 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 Ultravox Realtime 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. Reliability, 52 out of 100, up to 9.6 more on the total

Why it scored 52: A status page sits at status.ultravox.ai, but its robots rules blocked our reader on 1 and 2 October, robots.txt included, so we couldn't confirm components or history (10 of 20 for the page, 5 for unreadable history). Concurrency is published, 5 calls on pay as you go, no hard cap on Pro, and priority for up to 100 calls on Scale (15). Over-limit requests get 429, Scale accounts below their priority level get 503, and both carry a Retry-After header with exponential-backoff guidance. No idempotency guidance for call creation (12 of 15). No SLA found (0). The Realtime API is 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.

## 2. Security & auth, 46 out of 100, up to 9.5 more on the total

Why it scored 46: Plain API keys in the `X-API-Key` header, and each call returns a `joinUrl` so clients never see the key (20). We found no scoped or read-only keys (8 of 20). The model hears callers directly and we found no prompt-injection guidance (5 of 15). Call records and a delete-call API, but no audit log (8 of 15). No security.txt, and the privacy policy names no certification. Webhook signing is documented (5 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. Payments & pricing, 40 out of 100, up to 7.5 more on the total

Why it scored 40: No x402, MPP or L402 (0 of 40). $0.05 a minute, SIP at 0.5 cents and Pro at $100 a month are on the public pricing page (20). 30 free minutes with no card, per last week's check (20). Access starts with a human signup (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.

## 4. Agent ergonomics, 69 out of 100, up to 5 more on the total

Why it scored 69: Scored as an API, since there's no MCP server. Calls and messages list with cursors and page sizes (18 of 25). Paging and filters on calls (17 of 20). 429 and 503 come with Retry-After, which an agent can act on (14 of 20). No idempotency keys on call creation (5 of 20). A call needs only a system prompt, with default model and voice, and client SDKs for web, iOS, Android, Flutter, React Native and Python (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. Maintenance & community, 43 out of 100, up to 5 more on the total

Why it scored 43: The JavaScript client, ultravox-client 0.6.0, reached npm on 18 July 2026, 76 days before the check, with CI publishing and npm provenance. The newest platform change on the news and deprecation pages is the move to `ultravox-v0.7` on 22 December 2025 (20). One release in the last 90 days. An official example added on 28 and 29 September 2026 creates Inworld TTS 2 voices with pronunciation dictionaries through /api/voices, which no news entry announces, so we didn't count it (0). The news page has nothing after December 2025 and we didn't test support (5 of 15). Client SDKs for six platforms. JavaScript is current, the Python client's last release is 0.1.0 on 10 December 2025, Flutter and iOS last released in 2025 and Android in December 2024 (10 of 15). The JavaScript client runs CI and publishes from it, and the Python client needs 3.10 or later (8 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.

## 6. Schema & documentation, 84 out of 100, up to 2.6 more on the total

Why it scored 84: OpenAPI schema at api.ultravox.ai/api/schema/ (25). llms.txt with Markdown copies (10). Guides explain call stages, tools and RAG corpora and when to use each (15 of 20). The OpenAPI types call creation with enums and required fields (13 of 15). Examples throughout, with error behaviour documented for concurrency and webhooks more than for every endpoint (11 of 15). A deprecation guide with dated removals and model version strings to pin, but the news page's newest entry is 3 December 2025 (10 of 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, 75 out of 100, up to 2.2 more on the total

Made of editorial 67, provenance 82.

Why it scored 75: The Ultravox model code is MIT on GitHub, and the v0.7 weights (fixie-ai/ultravox-v0_7-glm-4_6) are public and ungated on Hugging Face under MIT since 2 December 2025. The hosted service is closed (26 of 30). The privacy policy (22 May 2025) says voice data and audio aren't used to train or fine-tune Ultravox's models, but keeps data for as long as the account exists, with call deletion through the API (15 of 30). A deprecation guide with a minimum 30-day window for pre-release capabilities and dated removals, such as the Qwen3 preview model on 22 December 2025 (18 of 20). Hosting in the United States is stated, with no subprocessor list found (8 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: ultravox.ai, registered 2023-10-05 (2 years) (7 of 15)
- security.txt: not found (0 of 10)

## 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.

- unchecked: status.ultravox.ai history. Its robots rules block our reader, robots.txt included
- Whether Ultravox holds SOC 2 or signs HIPAA BAAs. The privacy policy and the docs index name neither
- When Inworld TTS 2 voices and pronunciation dictionaries reached the API. An official example uses them, the news page doesn't say
- Whether the v0.7 Hugging Face repository holds the full model or only the audio adapter on top of GLM 4.6

## Weaknesses

- No MCP server
- News page, deprecation guide and Python client last updated in December 2025
- Plain API keys with no scopes
- Pay as you go caps at 5 concurrent calls
- No subprocessor list or stated certification

## 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.

- Honour `Retry-After` on 429 and 503 when creating calls
- Pin the model string (`ultravox-v0.7` or `ultravox-v0.6`) rather than relying on the default
- Hand clients the `joinUrl`, never the API key
- Delete calls through the API when you need data removed, there's no retention setting
- Use call stages for multi-step flows instead of one long prompt

## What the review panel asked for

- publish a Pro concurrency figure
- open the status page to readers
- scoped API keys
- a retention setting

## 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.