Fetch (reference) vs MarkItDown MCP (Microsoft)
Both are web search & scraping MCP servers, and we ran both through the same test (page read) in each app. Fetch (reference) passed 3 of 4 apps we tested it in; MarkItDown MCP (Microsoft) passed 0 of 4.
Test results side by side
| Fetch (reference) | MarkItDown MCP (Microsoft) | |
|---|---|---|
| Claude Code | Passed 2026-10-19 | Failed 2026-10-21 |
| Cursor | Passed 2026-10-19 | Failed 2026-10-21 |
| Claude Desktop | Passed 2026-10-14 | Failed 2026-10-14 |
| ChatGPT (Codex) | Failed 2026-10-20 | Failed 2026-10-21 |
| Tools exposed (protocol check) | 1 tools | 1 tools |
| GitHub stars | 91,266 (shared repo) | 190,251 (shared repo) |
| License | NOASSERTION | MIT |
| Last push | 2026-10-13 | 2026-10-04 |
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
Fetch (reference): Fetches a web page and returns it as Markdown so the agent can read it.
MarkItDown MCP (Microsoft): Microsoft's MarkItDown in MCP mode: converts files and office documents to Markdown your agent can read.
