MDN MCP (experimental) vs mcpdoc (LangChain)
Both are up-to-date docs for coding MCP servers, and we ran both through the same test (docs fact, mdn docs fact) in each app. MDN MCP (experimental) passed 3 of 3 apps we tested it in; mcpdoc (LangChain) passed 0 of 3.
Test results side by side
| MDN MCP (experimental) | mcpdoc (LangChain) | |
|---|---|---|
| Claude Code | Passed 2026-10-23 | Failed 2026-10-15 |
| Cursor | Passed 2026-10-23 | Failed 2026-10-15 |
| Claude Desktop | Not tested yet | Not tested yet |
| ChatGPT (Codex) | Passed 2026-10-23 | Failed 2026-10-22 |
| Tools exposed (protocol check) | 3 tools | — |
| GitHub stars | 192 | 1,031 |
| License | MPL-2.0 | MIT |
| Last push | 2026-10-21 | 2026-08-20 |
Results from our own harness runs, dated per cell. Stars and last push from GitHub at our last data refresh. We don’t pick a winner here: which one fits depends on your app and setup.
What each does
MDN MCP (experimental): MDN's experimental server for searching MDN docs and browser-compatibility data, so coding agents use current web platform facts. MDN says it logs queries while it's an experiment.
mcpdoc (LangChain): Serves llms.txt documentation files you choose to your coding agent, so it reads the real docs. The maker archived the repo.
