# Fix list: Extend API + MCP From Anchor Terminal's listing at https://www.anchorterminal.com/tools/extend, the October 2026 research run, assessed 1 October 2026. Grade B, 62.9 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 Extend API + MCP: 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, 43 out of 100, up to 10 more on the total Why it scored 43: REST calls use plain bearer keys, with test-mode keys hitting an isolated environment. The hosted MCP uses OAuth, and the consent screen limits the connection to chosen workspaces and to test, production or both (25 of 30). Test and production separation is the only least-privilege control we found, and the MCP includes write and delete tools with no documented confirmation (8 of 20). Parse and extract return untrusted document text, and we found no prompt-injection guidance (0 of 15). No audit log or per-call log documented (0 of 15). The trust centre at trust.extend.ai claims SOC 2 Type II, HIPAA and GDPR and gives security@extend.app. No security.txt and no bug bounty found (10 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, 30 out of 100, up to 8.8 more on the total Why it scored 30: No x402, MPP or L402 (0). Credit prices and credits per page are published without a login (20). Pay as you go starts with 10,000 free credits, and neither the pricing page nor the credits page says whether a card is needed (10 of 20, our call until that's confirmed). 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. ## 3. Reliability, 67 out of 100, up to 6.6 more on the total Why it scored 67: Statuspage at status.extend.ai with per-product components (20). Since 25 August the feed lists three incidents marked major and five minor. The longest major was 28 August, elevated latency and job processing for parsing and extraction from 17:05 to 18:41 UTC, the others were workflow run start failures on 25 August and five minutes of degraded API on 8 September. All five minor ones were latency (10). Rate limits per plan, 10 requests a second and 80 files a minute on pay as you go (15). 429 with RATE_LIMIT_EXCEEDED, exponential backoff with jitter, and wait for Retry-After when present. Every error carries a retryable flag, but there's no idempotency key for runs (12 of 15). Custom SLAs on Enterprise only, none published (0). Parse, Extract, Classify and Split 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. ## 4. Agent ergonomics, 68 out of 100, up to 5.2 more on the total Why it scored 68: The hosted MCP loads every tool by default, 86 per the 30 September check, which scores 5 for more than 30. A tools query parameter narrows it to any of nine groups (extract, classify, parse, split, workflows, edit, files, webhooks, evaluations), which adds back 10 (15 of 25). List and output controls not checked in detail (12 of 20). Every error says whether it's retryable and links to docs (20). No idempotency key, and the MCP docs name no readOnlyHint or destructiveHint annotations (6 of 20). A parse call needs only a file. Official SDKs in TypeScript and Python checked, Java and Go per the listing (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, 67 out of 100, up to 2.9 more on the total Made of editorial 48, provenance 86. Why it scored 67: Closed service, and the terms link is the website terms of use (15). Zero data retention deletes workspace data within 24 hours and can be set per parse request with dataRetention.mode, on request for all plans. AI subprocessors run under zero-retention, no-training terms. Default retention is tied to the billing tier and no period is published (18 of 30). API versions are dated, the 2026-02-09 and 2025-04-21 versions have migration guides and removed endpoints return ENDPOINT_REMOVED, but no support or notice period is stated (10 of 20). The trust centre names no subprocessors and no data locations. US and EU hosting come from the listing's 30 September check (5 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: extend.ai, registered 2017-12-15 (8 years) (11 of 15) - security.txt: not found (0 of 10) ## 6. Schema & documentation, 87 out of 100, up to 2.1 more on the total Why it scored 87: OpenAPI spec at docs.extend.ai/openapi/api-reference.json, per the 30 September check (25). llms.txt with a Markdown twin of every page (10). We couldn't read the MCP tool descriptions without an OAuth session. The API guides say what each processor is for and when to use the async endpoints (12 of 20). Typed request bodies in the spec, depth not re-checked (11 of 15). Error format with code, message, retryable, requestId and docUrl, seven general codes and domain codes on other pages (14 of 15). Dated API versions in the x-extend-api-version header, six of them back to 2024-02-01, plus a changelog of 38 entries (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. Maintenance & community, 82 out of 100, up to 1.6 more on the total Why it scored 82: TypeScript SDK v2.3.0 on 2026-09-25, 6 days ago (30). TypeScript v2.0.0, v2.1.0, v2.2.0 and v2.3.0 and Python v1.16.0 to v1.19.0 all since July (20). Public changelog, chat support on pay as you go and Slack on Scale, replies not sampled (10 of 15). Current official SDKs (15). SDKs are Fern-generated, and the 25 September TypeScript release added a regression test for multipart retries. CI not checked (7 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. ## 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: the MCP tool count. The docs don't state it and the tool list needs an OAuth session, so 86 comes from the 30 September check - unchecked: whether the 10,000 free credits need a card - unchecked: changelog entry dates, which didn't render for us - The listing priced Extract at 3 credits a page. The credits page says parse runs first and is billed on top, so Performance Extract is 5. Patched pricingNotes and unitPrices ## Weaknesses - Parse is billed on top of Extract, Split and Classify, so Performance Extract costs 5 credits a page - MCP loads every tool unless you pass ?tools=, including write and delete tools with no documented confirmation - Three major-impact incidents on the status page since 25 August - No security.txt, bug bounty, subprocessor list or published default retention period - No published support period for older API versions ## 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. - Connect with `https://mcp.extend.ai/mcp?tools=parse,extract,files` to keep the schema small - Every MCP call names a workspace and TEST or PRODUCTION; start in TEST - Read `retryable` on an error before retrying, and honour `Retry-After` on a 429 - Budget 5 credits a page for Performance Extract, since the parse step is billed on top - Sync `POST /extract` blocks up to 5 minutes; use `/extract_runs` and poll or a webhook for long files ## What the review panel asked for - Smaller default tool set - Publish tool descriptions - publish MCP tool descriptions ## 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.