Markdownify MCP Server
An MCP server that turns PDF, DOCX, XLSX, PPTX, images, audio, web pages and YouTube transcripts into Markdown
Install
git clone https://github.com/zcaceres/markdownify-mcp.git
cd markdownify-mcp
bun install
bun run buildBuild from source; requires bun and Python.
This is third-party code. Review the repository files before installing.
What it does
The server gives the agent a set of Markdown conversion tools built on markitdown. It handles PDFs, Word documents, Excel sheets, PowerPoint decks, images with metadata and audio with transcription. From the web it fetches pages, Bing search results and YouTube transcripts. Local file access can be limited to a list of allowed folders.
Who it is for. For people feeding the agent documents in various formats who want them as text.
Good fit when
- You need the agent to read a PDF or a slide deck
- You need a YouTube transcript for notes
- You need a web page as clean Markdown
Not a fit when
- You need exact layout and complex PDF tables without loss
- You cannot build from source and have no Docker
Example request
Convert report.pdf to Markdown and pull out the key findingsLimitations
There is no ready npm package: the project builds from source with bun and creates a Python environment with markitdown. The Docker image includes PDF support only, so audio and image recognition fail there. The YouTube and Bing tools call external services. File access is unrestricted by default.
How to disable. Remove the markdownify block from the MCP config and the cloned repository folder.
MCP
- Transport
- stdio
- Authentication
- not required
| Environment variables | |
|---|---|
| MD_ALLOWED_PATHS | Folders the server may read files from |
| MARKITDOWN_PATH | Path to the markitdown executable |
| REPOMIX_PATH | Path to the repomix executable |
Security check
- Reads local files with no folder limit by default
- Fetches pages and data from external services
README in short
The README lists supported formats and tools. Install is from source with bun, which prepares a Python environment, then the server is wired to the client via node and the build path. It documents environment variables for markitdown and repomix paths and access limits, plus Docker usage and its limits. MIT licensed.
FAQ
Why does audio transcription fail in Docker?
The image installs markitdown with PDF support only; audio and images need the local install.
How do I block access to other folders?
List allowed directories in MD_ALLOWED_PATHS.
Related
Microsoft's utility and MCP server that turn PDF, Word, Excel, PowerPoint, HTML and other files into Markdown for models
Anthropic skills
Skills
Anthropic's official skill examples: docx, pdf, pptx and xlsx documents, design, MCP builder and the Agent Skills spec
Anthropic's legal plugin suite: contracts, privacy, employment, IP, litigation, AI governance and legal education
A skill that builds editable PPTX decks from PDF, DOCX, URLs and Markdown with native shapes, charts and animations