Home / Compare / Context7 vs Windsurf Docs
Context7 vs Windsurf Docs MCP server
Verdict from our tests: Context7 passed more of our tasks (6 of 6) than Windsurf Docs (5 of 6). Both are in our documentation and docs search category.
| Context7 | Windsurf Docs | |
|---|---|---|
| Overall | Works | Partial |
| Tasks passed | 6 of 6 | 5 of 6 |
| Sign-in | None | None |
| Median task time | 204 ms | 740 ms |
| Slowest task | 1.7 s | 1.3 s |
| Gotchas | 1 | 2 |
| Date tested | Oct 16, 2026 | Oct 16, 2026 |
Context7: task by task
- Works connect (no auth) (439 ms)
- Works resolve library ID ("next.js") (1.3 s)
- Works query docs (/vercel/next.js generateMetadata) (1.7 s)
- Works resolve unknown library (204 ms)
- Works query unknown library ID (141 ms)
- Works error handling (malformed args) (34 ms)
Context7 gotchas
- Two calls are needed for most questions: resolve-library-id first, then query-docs with the returned ID.
Set up Context7: Claude | Claude Code | Cursor | VS Code | ChatGPT
Windsurf Docs: task by task
- Works connect (no auth) (753 ms)
- Works list tools (303 ms)
- Works search docs ("add an MCP server") (1.3 s)
- Works browse docs filesystem (ls /) (1.0 s)
- Partial search with nonsense query (740 ms)
- Works error handling (number as query) (200 ms)
Windsurf Docs gotchas
- One search returned about 137 KB. Very heavy on context.
- A nonsense query still returns unrelated docs instead of an empty result, so an agent cannot tell "nothing found" from a weak match.
Set up Windsurf Docs: Claude | Claude Code | Cursor | VS Code | ChatGPT
Best MCP servers for documentation and docs search | All tested connectors