Klavis-AI

MarkItDown MCP Server

Community Klavis-AI
Updated

Klavis AI (YC X25): MCP integration platforms that let AI agents use tools reliably at any scale

MarkItDown MCP Server

License: MIT

This project provides a Model Context Protocol (MCP) server that utilizes the markitdown library to convert various document types hosted at a given URL into Markdown format.

Features

  • Accepts a URL (HTTP/HTTPS) pointing to a document.
  • Supports conversion for:
    • PDF
    • Microsoft Office Documents (Word, PowerPoint, Excel)
    • HTML
    • Text-based formats (CSV, JSON, XML)
    • ZIP archives (extracts and converts contents)
    • EPub files
  • Handles optional authentication tokens for accessing protected resources.

Running Locally

There are two recommended ways to run this server locally: using Docker (recommended) or using a Python virtual environment.

Using Docker (Recommended)

This is the easiest way to get the server running without managing dependencies manually.

  1. Ensure Docker is installed on your system.
  2. Navigate to the root directory of the klavis project in your terminal.
  3. Build the Docker image:
    docker build -t markitdown-mcp -f mcp_servers/markitdown/Dockerfile .
    
    (You can replace markitdown-mcp with any tag you prefer)
  4. Run the Docker container:
    docker run -p 5000:5000 markitdown-mcp
    
    This command maps port 5000 on your host machine to port 5000 inside the container, where the server listens.

The server should now be running and accessible at http://localhost:5000.

Using Python Virtual Environment

This method requires Python 3.12 and pip to be installed.

  1. Navigate to this directory (mcp_servers/markitdown) in your terminal.
  2. Create a Python virtual environment:
    python3 -m venv .venv
    
  3. Activate the virtual environment:
    • On macOS/Linux:
      source .venv/bin/activate
      
    • On Windows:
      .\.venv\Scripts\activate
      
  4. Install the required dependencies:
    pip install -r requirements.txt
    
  5. Run the server:
    python server.py
    

The server should now be running and accessible at http://localhost:5000.

  1. Deactivate the virtual environment when you are finished:
    deactivate
    

MCP Server · Populars

MCP Server · New

    steveyegge

    beads-mcp

    Beads - A memory upgrade for your coding agent

    Community steveyegge
    mailtrap

    MCP Mailtrap Server

    Official mailtrap.io MCP server

    Community mailtrap
    statespace-tech

    ToolFront

    Data environments for AI agents

    Community statespace-tech
    PleasePrompto

    NotebookLM MCP Server

    MCP server for NotebookLM - Let your AI agents (Claude Code, Codex) research documentation directly with grounded, citation-backed answers from Gemini. Persistent auth, library management, cross-client sharing. Zero hallucinations, just your knowledge base.

    Community PleasePrompto
    pulsemcp

    Claude Code Agent MCP Server

    MCP (Model Context Protocol) Servers authored and maintained by the PulseMCP team. We build reliable servers thoughtfully designed specifically for MCP Client-powered workflows.

    Community pulsemcp