#5151st of 52 in MCP servers for developersFor data and documents
markdownify-mcp
MCP server that converts PDFs, Office files, audio and web pages to Markdown
- Stars
- 3k
- License
- MIT
- Last commit
- May 2026
- Last release
- May 2026
- Language
- TypeScript
2,476 on 31 Mar 20262,997 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Markdownify is a TypeScript MCP server that wraps Microsoft's markitdown to turn PDF, DOCX, XLSX, PPTX, images, audio, YouTube transcripts, Bing results and web pages into Markdown. Each format has its own tool, such as pdf-to-markdown or webpage-to-markdown. Setup uses bun, and a preinstall step creates a Python venv with markitdown[all].
Who it is for: Developers who want agents to read documents and web pages as Markdown
Strengths
- One tool per format: PDF, DOCX, XLSX, PPTX, image, audio, YouTube, Bing, web page
- MD_ALLOWED_PATHS restricts file-reading tools to listed directories
- Paths to markitdown and repomix can be overridden by environment variable
- Docker image available for the MCP catalog
Weaknesses
- Needs both Node/bun and a Python venv for markitdown
- Published Docker image installs markitdown[pdf] only; audio and image tools fail there
- File access is unrestricted unless MD_ALLOWED_PATHS is set
- Desktop config requires a manual clone, build and absolute path to dist/index.js
What it needs
- no GPU
- Docker
- Needs bun, Python, markitdown, repomix
Also in MCP servers for developers
See all 52| Rank | Project | Score |
|---|---|---|
| 1 | gitnexusCode knowledge graph exposed to coding agents through MCP tools | 92 out of 100 |
| 2 | codegraphLocal code knowledge graph that coding agents query over MCP | 82 out of 100 |
| 3 | n8n-mcpMCP server exposing n8n node docs, templates and workflow management to AI assistants | 77 out of 100 |
| #4 | serenaMCP server giving coding agents symbol-level code search, editing and refactoring | 74 out of 100 |
| #5 | chrome-devtools-mcpMCP server that lets coding agents drive and debug a live Chrome | 72 out of 100 |