Repolode
Submit
zcaceres

zcaceres/markdownify-mcp

An MCP server that converts PDFs, images, audio, Office files, YouTube transcripts and web pages into Markdown for AI assistants.

View on GitHub
Stars
2,997
Forks
255
Open issues
31
Last commit
26 days ago

What is it?

Markdownify is a Model Context Protocol server that gives AI assistants tools to turn files and web content into Markdown text. It covers PDFs, images, audio (with transcription), DOCX, XLSX and PPTX files, as well as YouTube video transcripts, Bing search results and general web pages. It can also read existing Markdown files.

The server is written in TypeScript and runs on Bun or Node. File conversion is delegated to a Python markitdown install, which the preinstall step sets up in a local virtual environment. Each conversion is exposed as a named tool, such as pdf-to-markdown or webpage-to-markdown, that an MCP client can call.

Access to local files can be restricted with the MD_ALLOWED_PATHS environment variable, which limits file-input tools to a list of directories. A Dockerfile is included, though the slim image omits audio and OCR support.

Key features

  • 01Converts PDF, DOCX, XLSX and PPTX files to Markdown
  • 02Transcribes audio files and extracts text from images
  • 03Fetches YouTube transcripts, Bing search results and web pages as Markdown
  • 04Exposes each conversion as a separate MCP tool
  • 05Restricts file access to allowed directories via MD_ALLOWED_PATHS

Use cases

  • Giving an AI assistant the contents of a PDF or Office document as text
  • Summarizing a YouTube video from its transcript
  • Feeding web pages or search results into an LLM as clean Markdown
  • Converting a meeting recording into a searchable transcript

Topics

#mcp#markdown#ai#pdf#ocr#transcription#web-scraping#typescript#anthropic#anthropic-ai

FAQ

What is markdownify-mcp?
An MCP server that converts PDFs, images, audio, Office files, YouTube transcripts and web pages into Markdown for AI assistants. Markdownify is a Model Context Protocol server that gives AI assistants tools to turn files and web content into Markdown text. It covers PDFs, images, audio (with transcription), DOCX, XLSX and PPTX files, as well as YouTube video transcripts, Bing search results and general web pages. It can also read existing Markdown files.
Is markdownify-mcp free to use? What license does it have?
markdownify-mcp is open source under the MIT license. Check the license terms before using it in commercial products.
Is markdownify-mcp still maintained?
Actively maintained. The last commit to markdownify-mcp was on Sep 25, 2026.
What can I use markdownify-mcp for?
Giving an AI assistant the contents of a PDF or Office document as text Summarizing a YouTube video from its transcript Feeding web pages or search results into an LLM as clean Markdown Converting a meeting recording into a searchable transcript
Where can I get markdownify-mcp?
The source code and installation instructions are on GitHub: https://github.com/zcaceres/markdownify-mcp

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects