Frequently Asked Questions
Everything you need to know about Context7 MCP Server
General Questions
What is Context7 MCP Server?
Context7 MCP Server is a documentation service that provides up-to-date library documentation to AI coding assistants. It connects via the Model Context Protocol (MCP) to tools like Cursor, VS Code, Claude Desktop, and Windsurf, giving them access to current documentation for over 9,000 libraries.
Who develops Context7?
Context7 is developed by Upstash, a company known for serverless data infrastructure including Upstash Redis and Upstash Kafka. Context7 leverages Upstash's edge infrastructure for fast, reliable documentation delivery.
What problem does Context7 solve?
Context7 solves the problem of AI hallucinations in coding assistance. AI models are trained on data with cutoff dates, meaning they do not know about library updates after their training. This leads to suggestions using deprecated patterns or non-existent APIs. Context7 provides current documentation so AI generates accurate, working code.
What is the Model Context Protocol (MCP)?
MCP is an open protocol that allows AI applications to communicate with external services. It standardizes how AI tools request and receive information, enabling seamless integration between different AI clients and various servers like Context7.
Pricing and Plans
Is Context7 free to use?
Yes, Context7 has a generous free tier that is suitable for most individual developers. You can use Context7 without an API key, though you will have lower rate limits. With a free API key from the dashboard, you get higher limits and usage tracking.
Do I need an API key?
Context7 works without an API key using default rate limits. However, we recommend getting a free API key for better performance, higher limits, and access to usage analytics.
Are there paid plans?
Yes, paid plans are available for teams and heavy users who need higher rate limits. Check the Context7 website for current pricing and plan details.
How do I get an API key?
Visit context7.com/dashboard, create a free account, and your API key will be displayed in the dashboard. Copy it and add it to your MCP client configuration.
Supported Clients and Libraries
Which AI coding assistants does Context7 support?
Context7 supports any MCP-compatible client. This includes Cursor, Visual Studio Code (with MCP extensions), Claude Desktop, Windsurf, and many others. As MCP adoption grows, more clients add support.
How many libraries does Context7 index?
Context7 indexes documentation for over 9,000 libraries and frameworks, covering most popular development tools including React, Vue, Next.js, Prisma, Tailwind CSS, and many more.
What if a library I need is not indexed?
You can request library additions through the Context7 GitHub repository or dashboard. Popular requests are prioritized for indexing. The library must have public documentation to be indexed.
How often is the documentation updated?
Context7's index is continuously updated. When libraries release new versions with documentation changes, Context7 typically indexes them within days. Critical libraries may be updated even faster.
Does Context7 support version-specific documentation?
Yes, Context7 tracks multiple versions of popular libraries. When you ask about a specific version, it provides documentation appropriate to that version.
Installation and Setup
How do I install Context7?
The fastest method is using the automatic installer: npx @upstash/context7-mcp@latest init. You can also specify a client: npx @upstash/context7-mcp@latest init --cursor. Manual installation involves adding Context7 to your MCP client's configuration file.
What are the prerequisites?
You need Node.js version 18 or higher installed on your system. Your AI client must support the Model Context Protocol. Most modern AI coding assistants do.
Do I need to update Context7 manually?
No, the recommended configuration uses @latest, which automatically uses the newest version each time it runs. You always get the latest features and bug fixes without manual updates.
Can I use Context7 with multiple editors?
Yes, you can configure Context7 in as many MCP clients as you want. Each client has its own configuration file, and they all connect to the same Context7 service.
Privacy and Security
Does Context7 access my code?
No, Context7 only provides documentation. It does not access, read, or transmit any of your code. The only data sent to Context7 is the library name and optional topic for documentation retrieval.
Is my usage data collected?
If you use an API key, basic usage metrics are collected for rate limiting and analytics (which libraries you query, when, and how often). Without an API key, only anonymized aggregate data is collected. See Context7's privacy policy for details.
Is Context7 secure to use in enterprise environments?
Yes, Context7 only retrieves public documentation and does not access your code or private systems. The connection uses HTTPS, and the service runs on Upstash's enterprise-grade infrastructure.
Technical Questions
How does Context7 work?
Context7 implements the MCP protocol with two main tools: resolve-library-id (converts library names to IDs) and get-library-docs (retrieves documentation). When your AI needs documentation, it calls these tools, and Context7 returns relevant sections from its indexed database.
What if Context7 is slow?
Context7 uses edge infrastructure for low latency. If you experience slowness, check your internet connection, try more specific queries (less data to retrieve), or verify the issue is with Context7 and not your AI client.
Does Context7 work offline?
No, Context7 requires an internet connection to fetch documentation from its servers. The documentation is not cached locally.
Can I self-host Context7?
Context7 is a hosted service by Upstash. Self-hosting is not currently available. However, you could build your own MCP server using the protocol specification if you have specific requirements.
Troubleshooting
Context7 is not working. What should I check?
First, verify your configuration file has valid JSON. Second, ensure Node.js is installed. Third, restart your AI client after configuration changes. Fourth, check for error messages in your client's console or logs.
I am getting rate limit errors. What do I do?
Get a free API key from context7.com/dashboard. The free tier without a key has conservative limits. With an API key, you get significantly higher limits.
My AI is not using Context7 documentation. Why?
Make sure you are asking about libraries by name. Generic questions like "how do I fetch data" may not trigger Context7. Be specific: "how do I fetch data with React Query".
The documentation seems outdated. Is this right?
Context7 updates regularly, but very new releases might not be indexed yet. If you notice outdated documentation for a popular library, report it through the Context7 GitHub repository.
Comparison Questions
How is Context7 different from web search?
Web search returns varied results including blog posts, Stack Overflow, and unofficial sources. Context7 specifically indexes official documentation, ensuring accuracy. It also provides semantically processed content optimized for AI context windows.
Should I use Context7 or just paste documentation manually?
Context7 automates what manual pasting does but better. It selects relevant sections, handles token limits, and works automatically for every query. Manual pasting is time-consuming and error-prone.
Can I use Context7 with my own documentation?
Context7 indexes public libraries. For private documentation, you would need to build your own MCP server or use a different RAG solution.