{
  "fixes": {
    "slug": "kagi-search",
    "name": "Kagi Search API",
    "listing": "https://www.anchorterminal.com/tools/kagi-search",
    "markdown": "# Fix list: Kagi Search API\n\nFrom Anchor Terminal's listing at https://www.anchorterminal.com/tools/kagi-search, the October 2026 research run, assessed 9 October 2026. Grade D, 48.7 out of 100.\n\nThis 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.\n\nFor a coding agent working on Kagi Search 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.\n\n## 1. Reliability, 8 out of 100, up to 18.4 more on the total\n\nWhy it scored 8: No status page is linked from the API docs, the pricing page, the API terms or the changelog, and we did not type an address for one (0). With no readable incident history the record takes the floor (5). No rate limit numbers are published. The terms say only that limits and quotas apply per plan (0). The OpenAPI file documents 429 as 'Rate limited or usage limit exhausted' with no `Retry-After` header or backoff guidance. The calls are read-only, and Kagi's own MCP server retries 429 and 5xx twice by default (3). The pricing page lists SLAs under Enterprise by quote, and the API terms say there is no uptime or service-level commitment unless agreed in writing (0). The API has been a public preview since 21 May 2026 (0). 8.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability):\n\nHosted APIs, MCP servers, models and platforms.\n\n- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 10, the surface agents use is generally available, not beta or preview.\n\nLocal packages, SDKs, frameworks and stdio MCP servers.\n\n- 20, installs from an official package with supported runtimes stated.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 15, version 1.0 or later, or declared stable.\n\nProtocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors.\n\n## 2. Payments \u0026 pricing, 20 out of 100, up to 10 more on the total\n\nWhy it scored 20: No x402, MPP or L402 (0). $12 per 1,000 searches and $4 per 1,000 extracted pages, published without a login (20). No free tier or trial for the API. The quick start asks for a payment method, and the $5 credit of 21 May 2026 was a one-off for subscribers (0). A person signs up in a browser and creates the key in the portal (0). 20.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-payments):\n\nThe published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/).\n\n- 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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API).\n\nPayment 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.\n\nOpen-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.\n\n## 3. Security \u0026 auth, 64 out of 100, up to 6.3 more on the total\n\nWhy it scored 64: Several named keys an account, each limited to chosen products and optionally to IP addresses or ranges, revoked by deleting. The v1 key travels only in the `Authorization` header. No expiry or OAuth (27). The API only reads. Per-key product limits and a spending cap in USD narrow what a leaked key can do (16). Search and Extract return third-party page text, and we found no prompt-injection guidance in the reference, the docs or the MCP README (2). The portal shows use count and last-used time per key and requests and cost per day per API, and every response carries a trace ID. No per-call log (7). security.txt with a contact, a bug bounty with a safe harbour policy, and an independent audit by Illumant cited on the security page, whose date we did not read. No SOC 2 or ISO 27001 statement found (12). 64.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-security):\n\n- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public.\n\nModels 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.\n\n## 4. Maintenance \u0026 community, 41 out of 100, up to 5.2 more on the total\n\nWhy it scored 41: The last dated API change is the Extract API fix in the changelog of 21 August 2026, 49 days before the check (20). That is the only dated API entry in the last 90 days. The MCP server's last commit is 7 July 2026 (0). A public changelog that credits user reports, developers@kagi.com, a Discord `#api` forum and GitHub discussions. We could not read the issue list, since the GitHub API answered 403 (9). Four generated clients last synced on 26 May 2026, Python tagged only as 1.0.0rc3, TypeScript not on npm, and `kagimcp` at 1.0.2 untagged. We did not look the server up in the MCP registry (8). The Python client has a generated CI workflow and the MCP server has a lock file but no CI workflow (4). 41.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance):\n\n- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.\n\nModels are read for deprecation notice periods and model churn rather than release counts.\n\n## 5. Schema \u0026 documentation, 71 out of 100, up to 4.7 more on the total\n\nWhy it scored 71: Public OpenAPI 3.0 file at kagi.com/api/docs/_spec/openapi.yaml, linked from the docs (25). No llms.txt on kagi.com or help.kagi.com, both 404. The docs' Markdown source is in a public GitHub repository, which is not the same as serving it to agents (0). Parameter descriptions say what each does and which option wins when lens, filters and query operators overlap, and the MCP tool says when to pick each workflow (14). Enums for `workflow`, `format`, `time_relative` and rule kinds, ranges on `page`, `limit`, `timeout` and `extract.count`, and one required field. Result `props` and `meta` are open objects (13). Three request examples with full responses and one 401 example. The error envelope is typed, but there is no list of error codes, and the file says its error documentation URLs return 404 for now (9). The path carries `/v1` and v0 is documented as legacy. The changelog is product-wide, with API items on 21 May, 16 June and 21 August 2026 (10). 71.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-schema):\n\nAPIs and MCP servers.\n\n- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 15, versioning and a public changelog.\n\nModels 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.\n\n## 6. Agent ergonomics, 75 out of 100, up to 4.1 more on the total\n\nWhy it scored 75: `limit` caps results, extraction of page text is opt-in, snippet length follows an account setting, and an experimental Markdown response format exists. The MCP server has two tools (22). `page` 1 to 10, `limit` 1 to 1,024, lens and `filters` for domains, keywords, file type, dates and region (20). Errors return a namespaced `code`, a message, the field `location` and a trace ID, but no code catalogue is published and the linked error pages are not live (12). Both operations are read-only, so a retry is safe, though the docs do not say so and failed-call billing is stated only for extraction. The MCP tools set no read-only annotations (10). Only `query` is required and defaults are stated. Generated clients exist for Python, TypeScript, Go and Rust, installable from GitHub. `@kagi/api` returned 404 on npm (11). 75.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics):\n\n- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 15, sensible defaults, few required parameters, and official SDKs in at least two languages.\n\nModels 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.\n\n## 7. Transparency \u0026 trust, 69 out of 100, up to 2.7 more on the total\n\nMade of editorial 54, provenance 83.\n\nWhy it scored 69: Closed service with clear API terms dated 22 September 2026. The MCP server is MIT and the OpenAPI file states Apache-2.0 (18). The privacy policy's API section agrees with the API terms. Queries are not stored against the account, load balancer and VM logs are kept 7 days and sampled Sentry errors 90 days. The Data Processing Addendum both documents cite is not yet published (23). v0 is marked deprecated with no end date, and the terms promise reasonable notice of breaking changes only where practicable (5). The privacy page names the payment processors and Sentry, the FAQ names GCP as the hosting provider and the docs name search sources, but there is no subprocessor list and no data locations (8). 54.\n\nThe checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency):\n\n- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted).\n\nThe other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two.\n\nProvenance checks not met in full (half of this category, computed from checked facts):\n\n- Terms of service: read, states 7 of the 7 things a reader expects, and has 2 clauses that cost points (6 of 10)\n- Privacy policy: read, states 4 of the 8 things a reader expects (7 of 10)\n- Status page: not found (0 of 10)\n\n## What we couldn't check\n\nWhat 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.\n\n- The lead gave `Authorization: Bot` and a GET request, as the help page example still shows. The OpenAPI file defines `POST /search` with a Bearer key. Whether v1 accepts both was not tested, since we send nothing to the API\n- unchecked: whether Kagi has a status page. None is linked from the pages we read, so it is scored as absent. A linked status page with a clean record would raise Reliability by up to 45\n- unchecked: the rendered API reference at kagi.com/api/docs, which is script-drawn. The OpenAPI file it is built from was read instead\n- unchecked: issues and discussions on kagisearch/kagi-docs and kagisearch/kagimcp. The GitHub API answered 403 to the issue list and we did not retry\n- unchecked: the MCP registry, PyPI download counts for `kagimcp`, and the tools the hosted server at mcp.kagi.com lists. The tool definitions were read from the source\n- unchecked: the date and scope of the Illumant security audit, which is on blog.kagi.com and was not read\n- The API terms forbid probing, scanning or testing Kagi's systems and reaching the API other than through the documented interface. This matters before any probe is run\n- The Data Processing Addendum cited by the API terms and the privacy policy is not yet published\n- The TypeScript client's README says `npm install @kagi/api@1.0.0`, and the registry returned 404 for that package on 9 October 2026\n\n## Weaknesses\n\n- Public preview since 21 May 2026. The API terms say it is under active development with no uptime or service-level commitment\n- No rate limit numbers are published, and no status page is linked from the docs, pricing, terms or changelog pages we read\n- No free allowance. The quick start asks for a Kagi account and a payment method before the first call\n- The API terms forbid storing or caching results beyond transient caching, building datasets from them, and probing or testing Kagi's systems\n- The help page example uses `GET` with `Authorization: Bot`, while the OpenAPI file defines `POST` with a Bearer key. The TypeScript client's README names an npm package that returned 404\n\n## What costs an agent a turn today\n\nThe notes we give agents before they call it. Each one is a workaround an agent shouldn't need.\n\n- Send `POST https://kagi.com/api/v1/search` with a JSON body and `Authorization: Bearer \u003ckey\u003e`, as the OpenAPI file defines. Only `query` is required\n- Set `limit` to cap results returned. In the mixed `search` workflow the MCP tool applies the cap to each result category separately\n- Use `extract.count` (1 to 10) only when page text is needed. Extraction is billed at the Extract rate on top of the search\n- Do not store results. The API terms allow only the transient caching an application needs, and require deletion on termination\n- On 429 back off before retrying. The response covers both rate limiting and an exhausted usage limit, and no `Retry-After` header is documented\n\n## When it's done\n\nSend 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.\n",
    "grade": "D",
    "score": 48.7,
    "assessed": "2026-10-09",
    "run": "October 2026 research run",
    "categories": [
      {
        "key": "reliability",
        "name": "Reliability",
        "score": 8,
        "maxGain": 18.4,
        "reason": "No status page is linked from the API docs, the pricing page, the API terms or the changelog, and we did not type an address for one (0). With no readable incident history the record takes the floor (5). No rate limit numbers are published. The terms say only that limits and quotas apply per plan (0). The OpenAPI file documents 429 as 'Rate limited or usage limit exhausted' with no `Retry-After` header or backoff guidance. The calls are read-only, and Kagi's own MCP server retries 429 and 5xx twice by default (3). The pricing page lists SLAs under Enterprise by quote, and the API terms say there is no uptime or service-level commitment unless agreed in writing (0). The API has been a public preview since 21 May 2026 (0). 8.",
        "checklist": [
          "Hosted APIs, MCP servers, models and platforms.",
          "- 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own).\n- 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.\n- 15, rate limits documented with numbers.\n- 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved.\n- 10, an SLA published for any paid tier.\n- 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.\n- 25, a public CI and test suite, passing on the default branch.\n- 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered).\n- 15, semver discipline and breaking changes called out in a changelog.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-reliability"
      },
      {
        "key": "payments",
        "name": "Payments \u0026 pricing",
        "score": 20,
        "maxGain": 10,
        "reason": "No x402, MPP or L402 (0). $12 per 1,000 searches and $4 per 1,000 extracted pages, published without a login (20). No free tier or trial for the API. The quick start asks for a payment method, and the $5 credit of 21 May 2026 was a one-off for subscribers (0). A person signs up in a browser and creates the key in the portal (0). 20.",
        "checklist": [
          "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.\n- 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.\n- 20, a free tier or trial that doesn't need a card.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-payments"
      },
      {
        "key": "security",
        "name": "Security \u0026 auth",
        "score": 64,
        "maxGain": 6.3,
        "reason": "Several named keys an account, each limited to chosen products and optionally to IP addresses or ranges, revoked by deleting. The v1 key travels only in the `Authorization` header. No expiry or OAuth (27). The API only reads. Per-key product limits and a spending cap in USD narrow what a leaked key can do (16). Search and Extract return third-party page text, and we found no prompt-injection guidance in the reference, the docs or the MCP README (2). The portal shows use count and last-used time per key and requests and cost per day per API, and every response carries a trace ID. No per-call log (7). security.txt with a contact, a bug bounty with a safe harbour policy, and an independent audit by Illumant cited on the security page, whose date we did not read. No SOC 2 or ISO 27001 statement found (12). 64.",
        "checklist": [
          "- 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.\n- 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions.\n- 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.\n- 0 to 15, audit logs or per-call visibility for the operator.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-security"
      },
      {
        "key": "maintenance",
        "name": "Maintenance \u0026 community",
        "score": 41,
        "maxGain": 5.2,
        "reason": "The last dated API change is the Extract API fix in the changelog of 21 August 2026, 49 days before the check (20). That is the only dated API entry in the last 90 days. The MCP server's last commit is 7 July 2026 (0). A public changelog that credits user reports, developers@kagi.com, a Discord `#api` forum and GitHub discussions. We could not read the issue list, since the GitHub API answered 403 (9). Four generated clients last synced on 26 May 2026, Python tagged only as 1.0.0rc3, TypeScript not on npm, and `kagimcp` at 1.0.2 untagged. We did not look the server up in the MCP registry (8). The Python client has a generated CI workflow and the MCP server has a lock file but no CI workflow (4). 41.",
        "checklist": [
          "- 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.\n- 20, at least three releases or dated changelog entries in the last 90 days.\n- 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.\n- 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models).\n- 10, package health, current dependencies and CI.",
          "Models are read for deprecation notice periods and model churn rather than release counts."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-maintenance"
      },
      {
        "key": "schema",
        "name": "Schema \u0026 documentation",
        "score": 71,
        "maxGain": 4.7,
        "reason": "Public OpenAPI 3.0 file at kagi.com/api/docs/_spec/openapi.yaml, linked from the docs (25). No llms.txt on kagi.com or help.kagi.com, both 404. The docs' Markdown source is in a public GitHub repository, which is not the same as serving it to agents (0). Parameter descriptions say what each does and which option wins when lens, filters and query operators overlap, and the MCP tool says when to pick each workflow (14). Enums for `workflow`, `format`, `time_relative` and rule kinds, ranges on `page`, `limit`, `timeout` and `extract.count`, and one required field. Result `props` and `meta` are open objects (13). Three request examples with full responses and one 401 example. The error envelope is typed, but there is no list of error codes, and the file says its error documentation URLs return 404 for now (9). The path carries `/v1` and v0 is documented as legacy. The changelog is product-wide, with API items on 21 May, 16 June and 21 August 2026 (10). 71.",
        "checklist": [
          "APIs and MCP servers.",
          "- 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool).\n- 10, llms.txt or Markdown docs served for agents.\n- 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.\n- 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs.\n- 0 to 15, examples and documented error responses.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-schema"
      },
      {
        "key": "ergonomics",
        "name": "Agent ergonomics",
        "score": 75,
        "maxGain": 4.1,
        "reason": "`limit` caps results, extraction of page text is opt-in, snippet length follows an account setting, and an experimental Markdown response format exists. The MCP server has two tools (22). `page` 1 to 10, `limit` 1 to 1,024, lens and `filters` for domains, keywords, file type, dates and region (20). Errors return a namespaced `code`, a message, the field `location` and a trace ID, but no code catalogue is published and the linked error pages are not live (12). Both operations are read-only, so a retry is safe, though the docs do not say so and failed-call billing is stated only for extraction. The MCP tools set no read-only annotations (10). Only `query` is required and defaults are stated. Generated clients exist for Python, TypeScript, Go and Rust, installable from GitHub. `@kagi/api` returned 404 on npm (11). 75.",
        "checklist": [
          "- 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).\n- 20, pagination, filtering and output-size controls.\n- 20, actionable, documented error responses, codes and messages an agent can recover from.\n- 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-ergonomics"
      },
      {
        "key": "transparency",
        "name": "Transparency \u0026 trust",
        "score": 69,
        "maxGain": 2.7,
        "reason": "Closed service with clear API terms dated 22 September 2026. The MCP server is MIT and the OpenAPI file states Apache-2.0 (18). The privacy policy's API section agrees with the API terms. Queries are not stored against the account, load balancer and VM logs are kept 7 days and sampled Sentry errors 90 days. The Data Processing Addendum both documents cite is not yet published (23). v0 is marked deprecated with no end date, and the terms promise reasonable notice of breaking changes only where practicable (5). The privacy page names the payment processors and Sentry, the FAQ names GCP as the hosting provider and the docs name search sources, but there is no subprocessor list and no data locations (8). 54.",
        "blend": "editorial 54, provenance 83",
        "checklist": [
          "- 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.\n- 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors).\n- 0 to 20, a deprecation policy or notices with dates.\n- 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."
        ],
        "checklistUrl": "https://www.anchorterminal.com/benchmark/#checklist-transparency"
      }
    ],
    "provenance": [
      {
        "label": "Terms of service",
        "value": "read, states 7 of the 7 things a reader expects, and has 2 clauses that cost points",
        "points": 6,
        "max": 10
      },
      {
        "label": "Privacy policy",
        "value": "read, states 4 of the 8 things a reader expects",
        "points": 7,
        "max": 10
      },
      {
        "label": "Status page",
        "value": "not found",
        "points": 0,
        "max": 10
      }
    ],
    "unchecked": [
      "The lead gave `Authorization: Bot` and a GET request, as the help page example still shows. The OpenAPI file defines `POST /search` with a Bearer key. Whether v1 accepts both was not tested, since we send nothing to the API",
      "unchecked: whether Kagi has a status page. None is linked from the pages we read, so it is scored as absent. A linked status page with a clean record would raise Reliability by up to 45",
      "unchecked: the rendered API reference at kagi.com/api/docs, which is script-drawn. The OpenAPI file it is built from was read instead",
      "unchecked: issues and discussions on kagisearch/kagi-docs and kagisearch/kagimcp. The GitHub API answered 403 to the issue list and we did not retry",
      "unchecked: the MCP registry, PyPI download counts for `kagimcp`, and the tools the hosted server at mcp.kagi.com lists. The tool definitions were read from the source",
      "unchecked: the date and scope of the Illumant security audit, which is on blog.kagi.com and was not read",
      "The API terms forbid probing, scanning or testing Kagi's systems and reaching the API other than through the documented interface. This matters before any probe is run",
      "The Data Processing Addendum cited by the API terms and the privacy policy is not yet published",
      "The TypeScript client's README says `npm install @kagi/api@1.0.0`, and the registry returned 404 for that package on 9 October 2026"
    ],
    "weaknesses": [
      "Public preview since 21 May 2026. The API terms say it is under active development with no uptime or service-level commitment",
      "No rate limit numbers are published, and no status page is linked from the docs, pricing, terms or changelog pages we read",
      "No free allowance. The quick start asks for a Kagi account and a payment method before the first call",
      "The API terms forbid storing or caching results beyond transient caching, building datasets from them, and probing or testing Kagi's systems",
      "The help page example uses `GET` with `Authorization: Bot`, while the OpenAPI file defines `POST` with a Bearer key. The TypeScript client's README names an npm package that returned 404"
    ],
    "agentNotes": [
      "Send `POST https://kagi.com/api/v1/search` with a JSON body and `Authorization: Bearer \u003ckey\u003e`, as the OpenAPI file defines. Only `query` is required",
      "Set `limit` to cap results returned. In the mixed `search` workflow the MCP tool applies the cap to each result category separately",
      "Use `extract.count` (1 to 10) only when page text is needed. Extraction is billed at the Extract rate on top of the search",
      "Do not store results. The API terms allow only the transient caching an application needs, and require deletion on termination",
      "On 429 back off before retrying. The response covers both rate limiting and an exhausted usage limit, and no `Retry-After` header is documented"
    ],
    "recheck": "https://www.anchorterminal.com/builders/#disputes"
  },
  "meta": {
    "attribution": "Anchor Terminal (https://www.anchorterminal.com)",
    "docs": "https://www.anchorterminal.com/docs/",
    "generatedAt": "2026-10-10",
    "license": "CC-BY-4.0",
    "method": "https://www.anchorterminal.com/benchmark/",
    "methodology": "0.4",
    "openapi": "https://www.anchorterminal.com/openapi.json",
    "preview": false,
    "run": "2026-10-01",
    "runLabel": "October 2026 research run"
  }
}