# Fix list: Azure AI Speech text-to-speech From Anchor Terminal's listing at https://www.anchorterminal.com/tools/azure-text-to-speech, the October 2026 research run, assessed 1 October 2026. Grade BB, 73.7 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 Azure AI Speech text-to-speech: 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-1M-character prices published without a login, though the page needs JavaScript and the Retail Prices API is the readable source (20). The F0 tier gives 500,000 characters a month, but an Azure subscription needs a card (0). 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. Schema & documentation, 65 out of 100, up to 5.7 more on the total Why it scored 65: Real-time synthesis takes SSML, a W3C format with documented Azure extensions, and we didn't confirm a public OpenAPI file for text-to-speech (10 of 25). No llms.txt found (0). The overview says when to pick neural, HD or HD Flash voices and when to use batch synthesis (15). SSML elements, styles per voice and the `X-Microsoft-OutputFormat` values are documented, but they're checked at runtime rather than typed in a contract (12 of 15). The REST page lists seven status codes with likely causes, and examples cover REST and the SDKs (13 of 15). Dated release notes and `api-version` values on batch synthesis (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. ## 3. Agent ergonomics, 75 out of 100, up to 4.1 more on the total Why it scored 75: API reading of the checklist. More than 40 output formats chosen by header, from 8 kHz telephony to 48 kHz, and synthesis events in the SDK (23 of 25). The voice list returns the region's voices in one response with no paging documented, while batch jobs list with paging, and per-request limits are published (10 minutes of audio, 50 voice or audio tags, 64 KB per WebSocket turn) (15 of 20). The REST page lists 400, 401, 415, 429, 502 and 503 with likely causes, and the SDK returns cancellation details with error codes (15 of 20). Retry guidance for 429, no idempotency key on batch jobs (10 of 20). Every REST request needs an SSML body and four headers, including the output format and a `User-Agent`. Speech SDKs in eight languages (12 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. ## 4. Reliability, 90 out of 100, up to 2 more on the total Why it scored 90: Azure status page with post-incident reviews (20). No review in the last 90 days names Speech. One on 29 September 2026 covers intermittent 5xx errors for Cognitive Services in Sweden Central from 10:03 to 15:58 UTC, which may touch Speech resources there, so we count it as minor. The public page only lists broad incidents (20). TTS quotas published, 20 transactions a minute on F0 and 30 a second on S0 by default, adjustable to 1,000, plus batch limits (15). The quotas page explains that 429 usually means backend capacity for a voice and region, and asks for retry logic, a gradual ramp and spreading load across regions (15). Covered by Microsoft's online services SLA (10). Neural and HD voices are GA. MAI-Voice-2-Flash is preview (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. ## 5. Security & auth, 90 out of 100, up to 1.8 more on the total Why it scored 90: Model reading of the checklist, with training and retention in place of least-privilege and injection lines. Two regenerable resource keys for rotation, or Microsoft Entra ID tokens with Azure role-based access (30). Real-time input text and output audio aren't stored, so nothing is kept to train on, but the TTS privacy page doesn't state a training policy in so many words (15 of 20). Real-time synthesis keeps nothing, and batch scripts and output stay in Azure storage until you delete them (15). Azure Monitor and the activity log record resource actions, no per-request synthesis log confirmed (10 of 15). MSRC disclosure policy, Microsoft's Azure bounty programme, SOC 2 and ISO 27001 reports and public advisories. The microsoft.com security.txt passed its Expires date on 2026-09-23 (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. ## 6. Maintenance & community, 80 out of 100, up to 1.8 more on the total Why it scored 80: Read as a model. Speech SDK 1.52 in September 2026 and MAI-Voice-2-Flash in public preview in July 2026 (30). Speech release notes for July, August and September 2026 (10). Retirements follow Microsoft's published lifecycle policy with dated notices (10). Public release notes and Microsoft Q&A, SDK issue tracker not checked (10 of 25). Speech SDK current in eight languages (15). The SDK is a closed binary, so we can't see its CI (5 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. ## 7. Transparency & trust, 88 out of 100, up to 1.1 more on the total Made of editorial 80, provenance 95. Why it scored 88: Closed service under Microsoft's product terms, with an MIT samples repository (15). The TTS privacy page, the privacy statement and the product terms agree on no storage for real-time synthesis and retention until deletion for batch (25 of 30). Dated retirements under Microsoft's lifecycle policy (20). Regions chosen per resource and a public sub-processor list (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): - security.txt: published but past its Expires date (5 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. - Whether a public OpenAPI file covers TTS batch synthesis or the voice list. We couldn't read the Azure REST specs tree. - Whether Microsoft states a no-training policy for TTS input in the product terms. The TTS privacy page doesn't say it directly. - Which date the listing's `lastRelease` of 2026-09-28 refers to. We found Speech SDK 1.52 in September and the last TTS service entry in July. ## Weaknesses - An Azure subscription needs a card, even for the free F0 tier - No llms.txt and no OpenAPI file for text-to-speech found - 429s often reflect busy capacity for a voice in a region, which a quota increase doesn't fix - The voice list comes back as one response per region with no paging documented - MAI-Voice-2-Flash, the low-latency model, is still in preview ## 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 SSML with `` and ``, and set `X-Microsoft-OutputFormat` and `User-Agent`. - On 429 retry with backoff, and try the voice's home region or another region rather than asking for more quota. - Keep each real-time request under 10 minutes of audio, or use batch synthesis. - Use Entra ID tokens instead of resource keys where the agent runs inside Azure. - Cache the voice list per region, since it returns hundreds of entries at once. ## What the review panel asked for - Publish prices as static text - Publish per-voice capacity guidance - Publish a time-to-first-audio figure ## 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.