DeepL API by DeepL

# Fix list: DeepL API

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/deepl-api, the October 2026 research run, assessed 1 October 2026. Grade BB, 72.4 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 DeepL 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: Status page at status.deepl.com, which the changelog says became one combined page with an API view on 16 July 2026 (20). The page renders only in a browser, and neither /history, the Statuspage JSON nor an RSS feed answered our reader, so we couldn't read its incidents. That's our limit, not a finding, but the rubric gives 5 for history we can't read (5). Request size limits are published (128 KiB a request, 16 KiB of headers) and plan character caps, but no request rate (5 of 15). 429 and 529 mean wait and resend, 456 means quota spent, and the docs say to retry 429 and 5xx with exponential backoff and to log the `X-Trace-ID` header. No Retry-After, and document uploads take no idempotency key (12 of 15). No SLA found for self-serve plans (0). The text and document APIs are GA (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. Payments & pricing, 30 out of 100, up to 8.8 more on the total

Why it scored 30: No x402, MPP or L402 (0). The plan structure is public, but the per-character rates sit on a pricing page that renders only in a browser and that we couldn't read, so plan-only credit (10). API Developer gives 1,000,000 characters in total with no card, per the 30 September check (20). A person signs up in a browser for a key, and the hosted MCP needs a DeepL account login (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.

## 3. Security & auth, 82 out of 100, up to 3.2 more on the total

Why it scored 82: Model reading of the checklist, with training and retention in place of the least-privilege and injection lines. Keys go only in the `Authorization` header since query and body auth were switched off in January and February 2026. API Key Permissions, GA on 24 June 2026, scope a key to chosen endpoints, the Admin API creates, labels and deactivates developer keys, and the hosted MCP signs in with OAuth (30). DeepL says paid-plan texts aren't used for training without consent, while free services may keep content for a limited time to train models. The privacy policy doesn't say which side API Developer keys fall on (15 of 20). Paid-plan texts are deleted once the job is done, document translation runs on DeepL's own infrastructure in the EEA (12 of 15). Usage reports by key and custom reporting tag, and a trace ID per call, but no audit log of key use found (10 of 15). security.txt PGP-signed and valid to 2026-12-01 per the 30 September check, a security@deepl.com address in the repositories and SOC 2 Type II claimed. No bug bounty found (15 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, 83 out of 100, up to 2.8 more on the total

Why it scored 83: API reading of the checklist. Responses carry the translation and detected language, with billed characters opt-in through `show_billed_characters`. The local MCP server has 13 compact tools, the hosted one up to 28 per DeepL's docs (22 of 25). Many strings per call up to 128 KiB, paging on style rules and translation memories, up to 5 glossaries a request (18 of 20). Errors carry a message, a code where available and a trace ID, and the docs say which codes to retry (20). Text translation is stateless with backoff guidance, but document jobs take no idempotency key and the MCP tools carry no readOnlyHint or destructiveHint annotations (10 of 20). Two required fields on translate, official SDKs for Python, Node.js and others, but a free key sent to the paid host fails (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, 84 out of 100, up to 1.4 more on the total

Made of editorial 67, provenance 100.

Why it scored 84: Closed service with published terms, the MCP server and OpenAPI repository under MIT (15). The privacy policy and data security page agree that paid-plan texts are deleted after the job and not used for training without consent, but neither places API Developer keys (22 of 30). Dated deprecations in the changelog, query and body auth deprecated in November 2025 and cut off on 20 January 2026 after a brownout, `/v2/languages` deprecated on 18 May 2026 (18 of 20). EEA processing for documents and regional endpoints for some customers, but we didn't find a public sub-processor list outside the trust centre at trust.deepl.com (12 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.

## 6. Maintenance & community, 87 out of 100, up to 1.1 more on the total

Why it scored 87: Changelog entries on 29 and 30 September 2026, and `deepl-mcp-server` 1.3.11 on 1 October (30). 16 dated changelog entries and 15 MCP server releases since 3 July (20). Public changelog, support centre and an issue triage workflow on the MCP repository, but we didn't read the GitHub issues (12 of 15). Official SDKs current, the MCP server pins `deepl-node` 1.28.0 (15). The MCP server runs lint, format, npm audit and tests on every push and nightly, with Renovate keeping dependencies current (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. Schema & documentation, 95 out of 100, up to 0.8 more on the total

Why it scored 95: OpenAPI 3 document with 43 paths, synced daily to the DeepLcom/openapi repository and served at developers.deepl.com (25). llms.txt with about 180 links, plus a keyless docs MCP server since 3 June 2026 (10). Operation descriptions say what each call does, and the docs say when to pick `latency_optimized` or `quality_optimized`, though not when to avoid an endpoint (15 of 20). Enums for formality, tag handling, model type and sentence splitting, with required fields marked (15). `/v2/translate` documents 11 response codes including 429, 456 and 529, with examples throughout (15). Versioned paths (v1, v2, v3) and a dated 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.

## 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: incident history on status.deepl.com, which renders only in a browser.
- unchecked: per-character rates for API Growth, shown only on a browser-rendered pricing page.
- The privacy policy names paid plans and free services but not API Developer keys, so whether API Developer text may be used for training is open.
- The patch moves `toolCount` from 8 to 13, the count in deepl-mcp-server 1.3.11.

## Weaknesses

- API Free and API Pro can't be bought any more, and API Developer's 1 million characters never reset
- Per-character rates only appear on a browser-rendered pricing page
- No published request rate and no SLA found for self-serve plans
- The privacy policy doesn't say how API Developer text is handled
- Office and PDF files bill at least 50,000 characters each

## 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.

- Match the host to the key. Keys ending `:fx` go to api-free.deepl.com, others to api.deepl.com
- Send the key in the `Authorization: DeepL-Auth-Key` header, since query and body keys now get 403
- Retry 429 and 529 with exponential backoff but stop on 456, which means the quota is spent
- Set `tag_handling` to html for markup so the tags survive and aren't billed
- Put a key-level usage limit on any key an agent holds, and scope it to the endpoints it needs

## What the review panel asked for

- Publish rates as plain text
- Restore a monthly free tier
- API Developer data terms

## 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.