The short version
- Choose Context7 if your agent mostly needs docs for well-known open-source libraries and you want it working in one command, with nothing to crawl.
- Choose SarvCrawl if you need docs Context7 does not have — internal wikis, vendor portals, PDFs, any site — searchable through the same MCP workflow.
SarvCrawl and Context7, side by side
| SarvCrawl | Context7 | |
|---|---|---|
| What it is | Crawl-to-knowledge-base API | Library docs service for coding agents |
| Best for | Content you want to search — web pages and documents together | Current docs for popular libraries inside your editor |
| Output | Markdown, HTML, JSON, links, images | Doc snippets and code examples as text |
| JavaScript rendering | Yes, where a page needs it | Not documented |
| Documents | PDF, Office, EPUB, images and audio; OCR for scans | Not documented |
| Chunking, embeddings, search | Built in (search on paid plans) | Built in: docs search over its catalogue |
| Anti-bot and proxies | Not the focus | Not applicable |
| Pricing | 1 billion free tokens, no card; paid plans | Free: 1,000 calls/mo, public libraries; Pro $10/seat/mo adds private repos |
| Hosting and licence | Hosted, or self-host: Apache-2.0 with an AGPL-3.0 engine | Hosted; the MCP server and CLI are MIT, the backend is closed |
| Integrations | REST, Node, Python, CLI, n8n, 21 MCP tools | MCP server, REST API, CLI, TypeScript SDK; Cursor, Claude Code, VS Code |
Where they really differ
Context7 keeps a shared, community-contributed catalogue of library documentation, indexed by version, and exposes it to coding agents through two MCP tools: resolve a library, then query its docs. For mainstream libraries it is close to zero setup, and that is a real strength — no crawling, no knowledge base to manage.
The catalogue is the boundary. Anyone can add a public library, mostly from a GitHub repository; private repositories need the Pro plan. The MCP server is open source, but the parsing, crawling and index behind it are hosted and closed.
SarvCrawl starts from any URL or file. Crawl a docs site, an internal handbook or a vendor portal, upload PDFs and Office files, and the result is your own knowledge base — chunked, embedded and searchable by keyword or meaning — that Claude Code, Cursor or VS Code reach through SarvCrawl's 21-tool MCP server.
Using both
Keep Context7 for mainstream libraries, and add SarvCrawl for the docs it does not cover — internal, vendor or niche — so your coding agent has both through MCP.
SarvCrawl vs Context7: questions
Is SarvCrawl a Context7 alternative?
For docs outside its catalogue, yes. For popular open-source libraries, Context7 is faster to set up, and the two run side by side in the same editor.
Can I add any docs site to Context7?
Anyone can submit a public library, mostly from a GitHub repository. Private repositories need a paid plan, and the index is shared and community-contributed.
How do I give a coding agent docs that Context7 lacks?
Crawl the docs site into a SarvCrawl knowledge base, then connect the SarvCrawl MCP server to Claude Code, Cursor or VS Code. The agent can search it like any other tool.
Is Context7 open source?
The MCP server, CLI and SDK are MIT. The API backend, parsing and crawling engines are private. SarvCrawl is Apache-2.0 with an AGPL-3.0 engine and can be self-hosted.