mdMCP
Turns Markdown files into local MCP servers, enabling users to define and run tools using simple Markdown syntax with zero dependencies.
README
mdMCP
Turn a Markdown file into a local MCP server, with zero dependencies.
pip install mdmcp
Quickstart
1. Write a tool file:
## Tool: health_check
Check if a URL is reachable.
### Parameters
- url (string, required): URL to check
### Execute
\```bash
STATUS=$(curl -sf -o /dev/null -w '%{http_code}' -m 5 "$MDMCP_URL")
echo "HTTP $STATUS"
\```
2. Serve it:
mdmcp serve tools.md
3. Connect to Claude Code:
{
"mcpServers": {
"my-tools": {
"command": "mdmcp",
"args": ["serve", "tools.md"]
}
}
}
That's it. Your Markdown is now an MCP server.
How it works
## Tool: namedefines a tool### Parametersdefines input schema (envvar-based, no string interpolation)### Executecontains a bash script inside a fenced code block- Parameters are passed as
$MDMCP_<NAME>environment variables
Commands
mdmcp list tools.md # List available tools
mdmcp run tools.md tool_name '{}' # Run a tool once
mdmcp serve tools.md # Start MCP stdio server
Security
- Minimal env: only PATH, HOME, USER, SHELL, LANG, TERM are inherited. API keys and tokens are NOT exposed to tool scripts.
- No string interpolation: parameters are passed as environment variables, not injected into shell code.
- Required param validation: missing required parameters return an error before execution.
- Timeout + process cleanup: 30s timeout with process group kill to prevent zombies.
Tips
- Always quote
"$MDMCP_VAR"in bash to prevent word splitting - Use
${MDMCP_VAR:-default}for optional params with bash-side defaults - You can use
python3 -cinside bash for complex logic (seeexamples/data-tools.md) - Use
set -euat the top of scripts for safer execution
Examples
See the examples/ directory:
- ops-tools.md — disk usage, port check, git log, docker status, health check
- dev-tools.md — find TODOs, run linter, check port (from Codex dogfood)
- data-tools.md — CSV summary, JSON inspect, file diff (from Gemini dogfood)
Known limitations
- Execute blocks only support
```fences (not~~~) - Parameter names must be
\w+(no spaces or special chars) - Nested
```inside heredocs will confuse the parser - All parameter values are strings in bash; cast manually if needed
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.