Sandbox
@zcaceres/markdownify-mcp

MCP server for file and web to Markdown conversion

Markdownify runs as an MCP server and exposes tools that convert common file types and web sources into Markdown. It uses `markitdown` and related helpers behind the scenes, so an agent can turn local files, web pages, and transcripts into readable text through tool calls.

2,990 stars254 forksTypeScriptUpdated 10d ago
Who it's for

Builders who want an MCP-enabled agent to convert files, web pages, and transcripts into Markdown.

What it delivers

You can hand your agent a PDF, image, audio file, or webpage and get Markdown back without manual copy and cleanup.

What it does

File to Markdown tools

Converts PDF, image, audio, DOCX, XLSX, and PPTX files into Markdown.

Web content tools

Turns YouTube transcripts, Bing search results, and general web pages into Markdown.

Markdown file retrieval

Reads existing `.md` and `.markdown` files through `get-markdown-file`.

Path restrictions

Uses `MD_ALLOWED_PATHS` or `MD_SHARE_DIR` to limit which local files the server can read.

Docker and desktop app setup

Runs locally with Bun or in Docker, and can be added to an MCP desktop app config.

How to get it

  1. 1Install dependencies
    bun install
  2. 2Build the project
    bun run build
  3. 3Start the server
    bun start
  4. 4Build and run
    docker build -t markdownify-mcp .
    docker run --rm -i \
      -v "$HOME/Documents:/data:ro" \
      -e MD_ALLOWED_PATHS=/data \
      markdownify-mcp

README

Markdownify MCP Server

markdownify mcp logo

Markdownify is a Model Context Protocol (MCP) server that converts various file types and web content to Markdown format. It provides a set of tools to transform PDFs, images, audio files, web pages, and more into easily readable and shareable Markdown text.

Markdownify Server MCP server

Features

  • Convert multiple file types to Markdown:
    • PDF
    • Images
    • Audio (with transcription)
    • DOCX
    • XLSX
    • PPTX
  • Convert web content to Markdown:
    • YouTube video transcripts
    • Bing search results
    • General web pages
  • Retrieve existing Markdown files

Getting Started

  1. Clone this repository

  2. Install dependencies:

    bun install
    

    The preinstall step creates a Python virtual environment at .venv and installs markitdown[all].

  3. Build the project:

    bun run build
    
  4. Start the server:

    bun start
    

Development

  • Use bun run dev to start the TypeScript compiler in watch mode
  • Modify src/server.ts to customize server behavior
  • Add or modify tools in src/tools.ts

Usage with Desktop App

To integrate this server with a desktop app, add the following to your app's server configuration:

{
  "mcpServers": {
    "markdownify": {
      "command": "node",
      "args": [
        "{ABSOLUTE PATH TO FILE HERE}/dist/index.js"
      ]
    }
  }
}

Environment variables

All paths default to sensible values; override only when the defaults don't fit your install layout.

VariableDefaultPurpose
MARKITDOWN_PATH<project>/.venv/bin/markitdown, then markitdown on PATHAbsolute path to the markitdown executable. Set this when you've installed markitdown system-wide (e.g. pipx install "markitdown[pdf]") instead of using the bundled venv.
REPOMIX_PATH<project>/node_modules/.bin/repomix, then repomix on PATHAbsolute path to the repomix executable used by git-repo-to-markdown.
MD_ALLOWED_PATHSunset (unrestricted)Path-delimiter-separated list (: on POSIX, ; on Windows) of directories the server is allowed to read. When set, all file-input tools (pdf-to-markdown, get-markdown-file, etc.) reject paths outside these directories.
MD_SHARE_DIRunsetDeprecated alias for MD_ALLOWED_PATHS (single directory). Still honored for backward compatibility.

Usage with Docker

Build and run:

docker build -t markdownify-mcp .
docker run --rm -i \
  -v "$HOME/Documents:/data:ro" \
  -e MD_ALLOWED_PATHS=/data \
  markdownify-mcp

Notes for the Docker MCP catalog (mcp/markdownify):

  • Mount any host directories you want the server to read into the container, then pass the container paths to the tools (e.g. /data/foo.pdf, not /Users/you/Documents/foo.pdf).
  • Set MD_ALLOWED_PATHS to the colon-separated list of mounted directories so the server enforces a read boundary that matches the bind mount.
  • The published Docker image installs markitdown[pdf] only — audio transcription and image OCR (audio-to-markdown, image-to-markdown) require the [all] extras and will fail in the slim image. Use the local install (bun install) for the full feature set.

Available Tools

  • youtube-to-markdown: Convert YouTube videos to Markdown

  • pdf-to-markdown: Convert PDF files to Markdown

  • bing-search-to-markdown: Convert Bing search results to Markdown

  • webpage-to-markdown: Convert web pages to Markdown

  • image-to-markdown: Convert images to Markdown with metadata

  • audio-to-markdown: Convert audio files to Markdown with transcription

  • docx-to-markdown: Convert DOCX files to Markdown

  • xlsx-to-markdown: Convert XLSX files to Markdown

  • pptx-to-markdown: Convert PPTX files to Markdown

  • get-markdown-file: Retrieve an existing Markdown file. File extension must end with: *.md, *.markdown.

    OPTIONAL: set MD_ALLOWED_PATHS to restrict every file-input tool to a list of directories, e.g. MD_ALLOWED_PATHS=/data/in:/data/out bun start.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Files in the repo

Repository payload19 top-level entries
  • .github
  • scripts
  • src
  • .dockerignore
  • .gitignore
  • .python-version
  • bun.lock
  • CLAUDE.md
  • Dockerfile
  • LICENSE
  • logo.jpg
  • package.json
  • preinstall.js
  • pyproject.toml
  • README.md
  • setup.bat
  • setup.sh
  • tsconfig.json
  • uv.lock

Discussion (0)

Ask about usage, or say what you built with it

Sign in to join the discussion.

No comments yet. Be the first to say what this is good for.

More connectors

Real-time global intelligence dashboard. AI-powered news aggregation, geopolitical monitoring, and infrastructure tracking in a unified situational awareness interface

86k

High-performance code intelligence MCP server. Indexes codebases into a persistent knowledge graph — average repo in milliseconds. 158 languages, sub-ms queries, 99% fewer tokens. Single static binary, zero dependencies.

43k

Universal provider proxy for OpenAI Codex & Claude Code — use any LLM (Claude, Gemini, Grok, DeepSeek, Ollama…) with Codex CLI, App, SDK, and Claude Code

14k
okf-memory/
okf-agent-memory

Git-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zero external databases or dependencies. Built in pure Go.

547
tirth8205/
code-review-graph

Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows.

31k
2akouwu/
reverify

Stop your AI from making things up — it proposes, deterministic tools decide, every claim checked against ground truth with evidence. Grounded facts and context survive resets. Reverse engineering is the proving ground. MCP server + CLI.

1.1k