Context7 by Upstash

# Fix list: Context7

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/context7, the October 2026 research run, assessed 1 October 2026. Grade BB, 73 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 Context7: 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, 68 out of 100, up to 6.4 more on the total

Why it scored 68: Scored as a hosted MCP server. Upstash runs an Atlassian Statuspage at status.upstash.com with Context7 and Context7 Console components (20). Its history lists no Context7 incident between July and September 2026, but Context7's own changelog for 4.0.1 and 4.0.2 describes "the 2026-08-11 mcp.context7.com outage", where 4.0.0 pushed concurrent upstream streams from about 10 to over 5,000 and the gateway returned 503 `reset reason: overflow`. We count that as one major outage on the vendor's own word (10). Monthly quotas are published (1,000 calls free, 2,000 per Pro seat) but the anonymous per-window limit is only described as low (8). 429 responses carry `Retry-After` and `RateLimit-*` headers, with backoff guidance and a status-code table in the API guide (15). The Enterprise plan lists an SLA but its terms aren't published, so half credit (5). The hosted endpoint is generally available (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. Security & auth, 69 out of 100, up to 5.4 more on the total

Why it scored 69: API keys with the `ctx7sk` prefix, hashed at rest and rotatable from the dashboard, and an OAuth endpoint at /mcp/oauth backed by Clerk, plus enterprise-managed auth. No per-key scopes (22). Every tool is read-only, so there's nothing destructive to confirm. Enterprise can restrict which libraries and source types a team can reach (15). Indexed documentation is third-party content. Context7 documents a two-pass prompt-injection and malware classifier on indexed content, which we can't test (12). Usage shows in the dashboard, and on-premise Enterprise adds an access audit log (8). SECURITY.md with private reporting and a 48-hour acknowledgement target, SOC 2 Type II through Upstash, no bug bounty found, no published advisories, and the supported-versions table still lists only 1.0.x while 4.1.1 ships (12).

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.

## 3. Agent ergonomics, 69 out of 100, up to 5 more on the total

Why it scored 69: Two tools. The definitions plus server instructions come to about 4,600 characters, roughly 1,200 tokens, most of it one description (20 of 25). Version 2.0.0 removed `page`, `limit` and the token budget, so there's no output-size control beyond writing a narrow query (5). Errors are readable and say what to do (a 429 names the dashboard or plans page, a 404 says to re-run `resolve-library-id`, a 401 names the `ctx7sk` key prefix), but they come back as ordinary text without `isError` (14). Both tools carry `readOnlyHint: true`, `destructiveHint: false` and `idempotentHint: true` (20). Two required parameters per tool and no optional ones. A server-side shim rewrites common hallucinated argument names. The SDK is TypeScript only, plus a REST API (10).

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. Payments & pricing, 60 out of 100, up to 5 more on the total

Why it scored 60: No x402, MPP or L402 (0). Overage is $5 per 1,000 calls on a public page (20). Free plan of 1,000 calls a month with no card, and anonymous calls work without any account (20). An agent can call the hosted endpoint with no key at all, so onboarding needs no human (20).

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.

## 5. Transparency & trust, 72 out of 100, up to 2.5 more on the total

Made of editorial 65, provenance 78.

Why it scored 72: The MCP server, CLI and SDK are MIT. The index and API behind them are closed, under Upstash's terms (25). The data-privacy page lists what is sent (query, library, client name and version, transport, an encrypted client IP), says model-written queries are stored anonymously for benchmarking with no retention period given, keeps API logs 30 days and deletes data within 30 days of a request (22). No deprecation policy. Breaking changes are flagged in the changelog at release, not in advance (3). Reranking subprocessors named (OpenAI, Google Gemini, Anthropic) and storage in the US and EU stated. The client analytics headers are disclosed, with no opt-out in the server (15).

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: context7.com, registered 2025-01-23 (1 year) (3 of 15)
- security.txt: not found (0 of 10)

## 6. Schema & documentation, 86 out of 100, up to 2.3 more on the total

Why it scored 86: Both tools take typed JSON Schema inputs built with Zod (25). context7.com/llms.txt and a docs index for LLMs exist, and docs/openapi.json is public (10). The server instructions say when to use it and when not to (refactoring, code review, general concepts), and `query-docs` says what comes first. The 2,006-character `resolve-library-id` description spends a third of its length telling the model how to format its own reply, which isn't tool guidance (14). Inputs are two required strings per tool with no enums or bounds, though the parameter descriptions give id formats (9). Good and bad query examples sit in the parameter text, and the API guide tables every status code with what to do (13). Semver releases with a changesets CHANGELOG that calls out breaking changes (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, 95 out of 100, up to 0.4 more on the total

Why it scored 95: @upstash/context7-mcp 4.1.1 on 14 September 2026, 17 days before the run date. ctx7 0.5.12 and the SDK 0.5.0 followed on 23 September (30). Thirteen MCP releases since 3 July, from 3.2.3 to 4.1.1 (20). 28 open issues on the run date, the oldest visible from 2 September, and pull requests merge most days. We couldn't read reply times (20). Listed in the official MCP registry as io.github.upstash/context7 4.1.1 (15). Dependabot, a test workflow on every push and pull request, and current dependencies (MCP SDK 2.0.0, undici 7) (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.

- How long the 11 August 2026 outage lasted and why it isn't on status.upstash.com
- unchecked: the numeric anonymous rate limit (docs say only "low")
- unchecked: maintainer reply times on issues (the issue list showed no comment counts)
- unchecked: whether failed or 429 calls count against the monthly quota
- How long model-written queries stored for benchmarking are kept

## Weaknesses

- The `resolve-library-id` description runs to 2,006 characters and includes reply-formatting instructions for the model
- No page, limit or token parameter since 2.0.0, so response size depends on how narrow the query is
- The 11 August 2026 mcp.context7.com outage (503s from a stream leak in 4.0.0) appears in the changelog but not on status.upstash.com
- Errors return as plain text without `isError`, so a client can't tell a 429 from a result
- SECURITY.md still lists only 1.0.x as supported while 4.1.1 ships

## 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.

- Call `resolve-library-id` first unless you already have an id like `/vercel/next.js`; `query-docs` answers a bare name with a not-found message
- Ask one concept per `query-docs` call; there's no size parameter, so a broad query returns a large chunk
- Read the text of every result, because a 429 or 404 comes back as normal content, not an error
- Send `Authorization: Bearer <ctx7sk key>` before looping; anonymous limits are low and unpublished
- Keep secrets and proprietary code out of `query`; queries are stored and sent to third-party models for reranking

## What the review panel asked for

- cut the resolve-library-id description
- set isError on failures
- retention period for stored queries
- current SECURITY.md versions

## 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.