DocuMCP

What it does

DocuMCP connects your AI assistant to your codebase and documentation pipeline. Give it access to a repository and it will deeply analyze the project structure, recommend the ideal static site generator (Jekyll, Hugo, Docusaurus, MkDocs, or Eleventy), generate Diataxis-structured documentation, and deploy everything automatically to GitHub Pages — all via natural language prompts.

Tools

• analyze_repository — Deep multi-layered analysis of project structure, dependencies, and documentation gaps
• recommend_ssg — Data-driven SSG recommendation with learning from past deployments and user preferences
• detect_gaps — Identifies missing documentation sections across tutorials, how-to guides, reference, and explanations
• deploy_pages — Automated GitHub Pages deployment with full outcome tracking and SSG-specific optimizations
• verify_deployment — Comprehensive post-deploy validation and troubleshooting
• analyze_deployments — Analytics and trend insights from deployment history
• manage_preferences — Store and recall user preferences for personalized recommendations
• manage_freshness — Track and validate documentation staleness with configurable thresholds
• manage_sitemap — Generate, validate, and maintain sitemap.xml for SEO

30+ tools total across analysis, generation, deployment, and maintenance.

Installation

Claude Desktop (claude_desktop_config.json):

{
  "mcpServers": {
    "documcp": {
      "command": "npx",
      "args": ["documcp"]
    }
  }
}

VS Code Copilot (settings.json):

{
  "mcp.servers": {
    "documcp": {
      "command": "npx",
      "args": ["documcp"]
    }
  }
}

Cursor (~/.cursor/mcp.json):

{
  "mcpServers": {
    "documcp": {
      "command": "npx",
      "args": ["documcp"]
    }
  }
}

Supported hosts

Claude Desktop, VS Code Copilot, Cursor, Gemini CLI.