Markdownify MCP is a Model Context Protocol server that converts various file types and web content to Markdown format.
The tool addresses the need to extract and standardize content from diverse sources into a unified, readable format. It wraps the markitdown Python library and exposes its conversion capabilities through the MCP interface, allowing Claude and other AI systems to transform PDFs, images, audio files, Office documents, and web pages into Markdown. For web content, it can fetch YouTube transcripts, Bing search results, and general web pages. It also retrieves existing Markdown files and can convert entire Git repositories to Markdown using repomix.
Developers should choose this tool if they are building Claude integrations or other MCP-compatible applications that need to ingest and normalize diverse content types. It suits projects where Claude needs to analyze PDFs, images, or web pages as part of its workflow. The tool is particularly valuable for scenarios involving document processing, research aggregation, or content extraction. The Docker deployment option makes it straightforward to run in containerized environments, though the full feature set including audio transcription and image OCR requires the complete markitdown installation with all extras; the published Docker image installs only the PDF-specific dependencies.
The project maintains active development with TypeScript compilation in watch mode available for local iteration. The codebase is organized around a central server file and a tools module, making it accessible for customization. Environment variables allow flexible configuration of executable paths and read-access boundaries, with the latter enabling security controls when the server runs in restricted environments. The tool supports both local installation and Docker deployment, with clear documentation of the differences in capability between the two approaches.
