Document conversion between formats like PDF, Word, Markdown, and HTML is a common workflow need. MCP servers provide powerful conversion capabilities that let Claude transform documents while preserving structure and formatting. Here are the best options available.
Recommended MCP Servers
MarkItDown
microsoft/markitdown
Official Microsoft library that converts PDF, Word, Excel, PowerPoint and more to Markdown for LLM consumption.
Key Features
- Converts PDF, DOCX, XLSX, PPTX to Markdown
- Preserves document structure
- Supports images with OCR
- Official Microsoft support
Installation
pip install 'markitdown[all]'MarkItDown is the gold standard for document-to-Markdown conversion. It handles virtually every common format, preserves structure like headings and tables, and is officially maintained by Microsoft. The go-to choice for most conversion needs.
MCP Pandoc
vivekVells/mcp-pandoc
Bidirectional document conversion using Pandoc, supporting Markdown, HTML, PDF, DOCX, EPUB, and LaTeX.
Key Features
- Bidirectional conversion
- PDF and DOCX output
- EPUB and LaTeX support
- Reference document styling
Installation
pip install mcp-pandocMCP Pandoc is the choice when you need to create documents, not just read them. Convert Markdown to PDF, generate Word documents, or create EPUB books. Pandoc's format coverage is unmatched.
Markdownify MCP
zcaceres/markdownify-mcp
Convert PDFs, images, audio, DOCX, XLSX, PPTX, and web content to Markdown.
Key Features
- Multi-format to Markdown
- Image OCR support
- Audio transcription
- YouTube transcript extraction
Installation
pnpm install && pnpm run buildMarkdownify goes beyond documents to handle images, audio, and video content. If your workflow involves converting diverse media types to text, this all-in-one solution covers everything.
Fetch MCP (Anthropic)
anthropics/mcp-server-fetch
Official server for fetching web content and converting HTML pages to Markdown.
Key Features
- Web page to Markdown
- Clean content extraction
- Official Anthropic support
- Handles most websites
Installation
npx @modelcontextprotocol/server-fetchThe official Fetch MCP cleanly converts web pages to Markdown. When you need to bring web content into your workflow in a format Claude can work with, this is the reliable choice.
Quick Comparison
| MCP | Stars | Best For | Pros | Cons |
|---|---|---|---|---|
| MarkItDown | >80k | Docs to Markdown | Best format coverage | Read-only |
| MCP Pandoc | >400 | Bidirectional | Creates PDFs/DOCX | Requires Pandoc |
| Markdownify | >2k | Mixed media | Audio/video support | More setup |
| Fetch MCP | >2k | Web to Markdown | Official support | Web only |
Getting Started
To convert documents with Claude Code:
For reading documents, install MarkItDown:
pip install 'markitdown[all]'For creating PDFs/DOCX, install MCP Pandoc:
pip install mcp-pandocFor web pages, use the official Fetch MCP.
Example prompt: "Convert this Word document to Markdown and clean up the formatting."