Markdownify MCP

Add to your client

Copy the config for your MCP client and paste it into its config file.

{
  "mcpServers": {
    "markdownify-mcp": {
      "command": "node",
      "args": [
        "/absolute/path/to/markdownify-mcp/dist/index.js"
      ]
    }
  }
}

Before you start

• Node.js with a JavaScript runtime/package manager — the repo uses Bun (bun install / bun run build / bun start)
• Python 3 available on the system — the install step creates a .venv and installs markitdown[all] (audio/image conversion may need additional system tools like ffmpeg/exiftool depending on input)
• No API keys or credentials required for local file conversion; web/search tools fetch public URLs

About Markdownify MCP

Tools & capabilities (10)

When to use it

• Use it when you need to feed a PDF, Word, Excel, or PowerPoint file into an LLM as clean Markdown context
• Use it when extracting text from images via OCR or transcribing audio into readable Markdown
• Use it when scraping a web page or YouTube transcript into Markdown for summarization or RAG
• Use it when you want conversions to happen locally so sensitive documents never leave your machine
• Use it when building a document-ingestion pipeline and want a single MCP for many input formats
• Use it when you need to lock down which directories the agent can read by setting MD_ALLOWED_PATHS

Quick setup

1. 1Clone the repo: git clone https://github.com/zcaceres/markdownify-mcp
2. 2Install dependencies: bun install (this creates a .venv and installs markitdown[all])
3. 3Build the server: bun run build
4. 4Add it to your MCP client config as a command server: node /absolute/path/to/dist/index.js
5. 5Optionally set MARKITDOWN_PATH, REPOMIX_PATH, and MD_ALLOWED_PATHS environment variables
6. 6Restart the client and test pdf-to-markdown on a sample file to verify

Markdownify MCP FAQ