# Fix list: DeepSeek Search and Fetch API From Anchor Terminal's listing at https://www.anchorterminal.com/tools/deepseek-search-api, the October 2026 research run, assessed 17 October 2026. Grade D, 53.2 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 DeepSeek Search and Fetch 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. Security & auth, 38 out of 100, up to 10.9 more on the total Why it scored 38: The ordinary DeepSeek bearer key, revocable, with no scopes, so a key used for search can also spend on models. No secret in a URL (20 of 30). Both endpoints only read, so there's nothing destructive to approve, but no search-only key (15 of 20). Results and fetched pages are untrusted web text. The change log says the engine filters 'SEO spam, GEO manipulation, and low-quality AIGC', with no technical detail and no prompt-injection guidance (3 of 15). No per-call log or audit log found in the docs. The platform console wasn't readable (0 of 15). No security.txt on deepseek.com or www.deepseek.com, and no disclosure policy, bug bounty or certification found (0 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. ## 2. Payments & pricing, 20 out of 100, up to 10 more on the total Why it scored 20: No machine payment protocol. Unauthenticated POSTs to both endpoints returned 401 with no 402 challenge (0). $0.75 per 1,000 successful requests for each endpoint, published without a login on the guide and the pricing page (20). No free tier for direct calls. The free period to 31 October 2026 covers only server-side calls made through DeepSeek Harness, Claude Code or Codex, which still bill model tokens (0). Browser sign-up and a prepaid top-up (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. Schema & documentation, 55 out of 100, up to 7.3 more on the total Why it scored 55: No OpenAPI file found. The API reference is rendered from a schema, but no spec is linked, and /openapi.json and /openapi.yaml serve the docs page (0). No llms.txt, since api-docs.deepseek.com answers /llms.txt with HTML (0). The guide says what each endpoint is for and which result field suits a UI or a model, and gives the score thresholds for quality and authority, but not when to use Search content against a Fetch (14 of 20). Types, required fields, defaults and limits on every field, such as `query` 1 to 128 characters and `count` 1 to 50. Filters live inside the free-text query (13 of 15). curl and Python examples, an error body and a table of ten status codes with retry guidance. The reference shows only the 200 response (13 of 15). Dated change log and a /v1 path (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. ## 4. Reliability, 65 out of 100, up to 7 more on the total Why it scored 65: Hosted reading. status.deepseek.com, run on Flashduty, has component history, but the endpoints have no component of their own. The nearest is Search Service, grouped under Chat Service (10 of 20). The change log says the same in-house engine runs web and API search, so we read that component's record. It shows three incidents in the 90 days to 17 October, Search service errors of 40 minutes on 29 July, 6 on 14 August and 39 on 18 August. That's minor only, scored for the engine since the API opened today (20 of 30). Per-account limits of 5 requests a second for each endpoint (15). The guide says to wait and retry on 408, 429 and 5xx, with no Retry-After or backoff schedule, and both endpoints are read-only (10 of 15). No SLA found (0). Opened to everyone with no beta label (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. Maintenance & community, 42 out of 100, up to 5.1 more on the total Why it scored 42: Closed service. Opened on 17 October 2026 (30). The change log has several entries in 90 days, but the 17 October one is the only one about Search or Fetch (0 of 20). Dated change log, api-service@deepseek.com and a Discord server, not tested for replies (12 of 15). No official SDK (0 of 15). No SDK or package to assess (0 of 10). Low because the API is a day old, by the checklist. 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. Transparency & trust, 60 out of 100, up to 3.5 more on the total Made of editorial 35, provenance 84. Why it scored 60: Closed service with published terms under PRC law, which assign Outputs to the customer (15). The privacy policy of 10 February 2026 says personal data trains models unless you opt out by email and gives no retention period. It also says search runs on 'third-party APIs' that receive input keywords, which disagrees with the in-house engine described on 17 October, and nothing covers fetched URLs (5 of 30). No deprecation policy, and the only dated notice is the end of free server-side calls (5 of 20). Data location disclosed as the PRC, third parties named only by category (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): - Terms of service: published, but our reader couldn't read it (7 of 10) - Privacy policy: published, but our reader couldn't read it (7 of 10) - security.txt: not found (0 of 10) ## 7. Agent ergonomics, 81 out of 100, up to 3.1 more on the total Why it scored 81: Responses can be sized with `count` and `need_content`, off by default, and each result carries both a short snippet and a 600 to 800 character summary. No field selection, and full content has no length cap stated (20 of 25). Site, language, date and phrase filters, up to 50 results, no pagination (15 of 20). Ten documented codes with what to do, 402 for an empty balance and 404 for an unfetchable page. `param` is always null and `code` repeats the type (17 of 20). Both endpoints are documented as read-only and safe to retry, and failed or empty calls aren't billed (20). Only `query` or `url` is required. No official SDKs, the docs use raw HTTP and Python requests (9 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. ## 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: per-call usage or logs in the console. platform.deepseek.com answered with a 202 holding page and wasn't read - unchecked: the number for Fetch's per-domain limit, which the guide mentions but doesn't give - Whether the Search Service component on the status page covers the API endpoints or only search in the chat app - Whether the privacy policy's statement that search runs on third-party APIs still applies, given the 17 October change log describes an in-house engine - Whether search queries and fetched URLs are kept or used for training. The privacy policy covers Inputs generally and says nothing specific - How a granted balance is given, which the pricing page says is spent first ## Weaknesses - Opened on 17 October 2026, with no service record of its own and no component for it on the status page - 5 requests a second per account for each endpoint, plus an unpublished per-domain limit on Fetch - No OpenAPI file, no llms.txt, no official SDKs and no free tier for direct calls - One unscoped key for search and models, no security.txt, and no prompt-injection guidance for returned page text - Data stored in the PRC, used for training unless the customer opts out by email, and no retention period stated ## 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. - Keep `query` to 128 characters and put filters in it as operators, such as `site:github.com lang:en after:2026-01-01` - Leave `need_content` false and read `summary` first, then call Fetch on the URLs you keep - Stay under 5 requests a second per account across all keys. On 408, 429 or 5xx wait and retry, since no Retry-After is documented - Don't retry a Fetch 403 or 404 unchanged. The page is blocked by policy, unreachable or has no Markdown - Don't send private queries or an end user's IP in `localization`. Data is stored in the PRC ## What the review panel asked for - Scoped or capped keys - Free allowance for direct calls - Publish an OpenAPI file - Document error responses in the reference - reconcile the privacy policy with the engine - typed filter fields - search-only key scope - state retention for queries ## 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.