Context7 Discovery

AI coding assistants — Claude Code, Cursor, VS Code Copilot, and others — help developers discover tools while they work. Context7 is the bridge between those assistants and your App.

When you enable Context7 on your server, developers using AI assistants can find your server by asking questions like "what tools are available for web scraping?" or "find me an App for GitHub analytics." Your server's documentation surfaces directly in the conversation, alongside connection instructions.

What Context7 discovery does for you

When your server is listed on Context7:

• AI coding assistants can recommend your server when developers search for tools you provide
• Documentation appears in-context — not as a link to click, but as content the AI can use to explain what your server does
• Subscribe instructions are included — developers see how to connect without leaving their editor

This is passive, always-on discovery. Once enabled, your listing stays current automatically on a daily schedule — no redeploy needed.

Enable Context7 discovery

1. Open your server dashboard and select your server
2. Find the Context7 Discovery toggle on your server detail page
3. Turn it on

mctx queues a documentation submission for your server. The listing goes live on the next daily publishing run — AI coding assistants can start finding your server the following day.

Note: This feature is opt-in. Your server must have at least one live deployed version before documentation can be generated.

What gets published

mctx automatically builds a structured document from your server's information:

The document is served at https://your-slug.mctx.ai/llms.txt — a standard format that AI coding assistants understand. You can visit this URL at any time to see exactly what is published.

Your README content is the most influential field in this document — it is how AI assistants understand what your server does and when to recommend it. See Writing a good README for guidance on structuring it for both human readers and AI discovery.

mctx adds platform metadata automatically. When publishing to Context7, mctx injects a prefix and postfix around your README — this includes your server slug, subscribe link, connection endpoint, and version details. You do not need to include any mctx branding or hosting references in your README. Focus your README entirely on subscriber value and usage.

How Context7 ranks your server

Context7 uses a benchmark score to determine how prominently your server surfaces in AI assistant recommendations. Understanding this score helps you write content that ranks higher.

Benchmark score — how your documentation answers developer questions

Context7 auto-generates questions that developers might ask an AI assistant — things like "what does this server do?", "why would I use it?", and "how do I invoke X?" It then measures how well your documentation answers those questions. A README that clearly explains what the server does, why someone would use it, and how to invoke its capabilities will naturally score higher.

The recommended README structure (value and intent → usage in conversation → tool reference → example phrases → example responses) is specifically designed to answer these kinds of questions. Writing with that structure is the most direct way to improve your benchmark score.

Trust score — repository signals

Context7 also factors in GitHub repository signals when ranking servers:

• Stars — More stars indicate community interest and credibility
• Commit activity — Recent, consistent commits signal an actively maintained server
• Update recency — A README that has not been touched in months is a negative signal

An actively maintained repo with a current, descriptive README surfaces more prominently in AI assistant recommendations than a stale one. Regular README updates and active commit history keep your trust score healthy.

Automatic updates on a daily schedule

mctx refreshes your Context7 listing automatically on a daily schedule. You do not need to do anything — just keep the toggle on and the listing stays current.

This means:

• New tools you add appear in AI assistant responses on the next daily run
• An updated README flows through automatically
• Version numbers stay accurate

Disable Context7 discovery

You can turn off the toggle at any time from your server's detail page.

When you disable Context7 discovery:

• The /llms.txt endpoint on your server stops responding
• mctx stops refreshing your listing on future scheduled runs

One thing to know: Context7 does not provide a way for mctx to delete your listing. After you disable discovery, your listing remains in Context7 but stops receiving updates. It will gradually become stale. There is no way to force immediate removal.

If you need your server removed from Context7 entirely, contact Context7 directly at context7.com.

Next Steps

• Publish to MCP Community Registry — Official directory discovery
• Directories and Social — Expand your reach further
• Your Vanity Link — Share your server everywhere

See something wrong? Report it or suggest an improvement — your feedback helps make these docs better.