Mati MCP Server

MCP server that connects Cursor, Claude, and other AI tools to the Mati content API. Create drafts, schedule posts, and publish to Twitter—all from your AI assistant.

Install

From PyPI (once published):

pip install mati-mcp-server

From source (development):

cd mati-mcp-server
make install   # creates venv, installs deps
make run       # or: source venv/bin/activate && mati-mcp-server

Configure

Set these environment variables (or add to your shell profile):

Add to Cursor

1. Open Cursor → Settings → Features → MCP

2. Click Add New MCP Server

3. Configure:

   • Name: Mati (or any name)
   • Type: stdio
   • Command: mati-mcp-server

4. If you need to pass env vars, use a command like:

   env MATI_API_BASE_URL=https://api.dev.hiremati.ai MATI_API_KEY=your-key mati-mcp-server
   

5. Save and refresh

Add to Claude Desktop / Claude Code

claude mcp add mati mati-mcp-server

Or with env vars:

env MATI_API_KEY=your-key claude mcp add mati mati-mcp-server

Tools

• create_draft – Create or update a content draft
• list_drafts – List your drafts
• list_scheduled – List scheduled content
• update_draft – Update an existing draft
• schedule_post – Schedule content for publishing
• post_content – Post content to Twitter

All content is linked to your Mati user (via your API key).

Optional: Run as HTTP server

To run a standalone HTTP server (e.g. for remote connections):

mati-mcp-server --http

Server runs at http://localhost:8000/mcp.

Publishing to PyPI (for maintainers)

# 1. Install build tools
pip install build twine

# 2. Build the package (creates dist/*.whl and dist/*.tar.gz)
python -m build

# 3. Upload to PyPI (prompts for username/password or token)
twine upload dist/*