Skip to main content
Glama

DocuMCP - Intelligent Documentation Deployment MCP Server

CI CodeQL Coverage npm version Ask DeepWiki

DocuMCP is an intelligent Model Context Protocol (MCP) server that revolutionizes documentation deployment for open-source projects. It provides deep repository analysis, intelligent static site generator recommendations, and automated GitHub Pages deployment workflows.

TL;DR

DocuMCP analyzes your repository, recommends the perfect static site generator (Jekyll, Hugo, Docusaurus, MkDocs, or Eleventy), creates professional documentation structure following Diataxis principles, and deploys it automatically to GitHub Pages. Just say "analyze my repository and deploy documentation" to get started.

Related MCP server: Synapse

Features

Core Capabilities

  • πŸ” Repository Analysis: Deep multi-layered analysis of project structure, dependencies, and documentation needs

  • 🎯 SSG Recommendations: Data-driven recommendations for Jekyll, Hugo, Docusaurus, MkDocs, or Eleventy

  • πŸ“š Diataxis Framework: Automatic creation of well-structured documentation following proven principles

  • πŸš€ GitHub Pages Deployment: Automated workflow generation with SSG-specific optimizations

  • βœ… Deployment Verification: Comprehensive checks and troubleshooting for successful deployments

Intelligence & Learning (Phase 2)

  • 🧠 Historical Intelligence: Learns from past deployments to improve recommendations

  • πŸ‘€ User Preferences: Personalized recommendations based on your preferences and patterns

  • πŸ“Š Deployment Analytics: Comprehensive insights into deployment patterns and success rates

  • 🎯 Smart Scoring: Intelligent SSG scoring based on success rates from similar projects

  • πŸ“ˆ Trend Analysis: Identifies deployment trends and provides health scores

Documentation Maintenance (v0.5.2+)

  • πŸ“… Freshness Tracking: Monitor documentation staleness with configurable thresholds

  • βœ… Freshness Validation: Initialize and update freshness metadata automatically

  • πŸ—ΊοΈ Sitemap Management: Generate, validate, and manage sitemap.xml for SEO

  • πŸ”— Knowledge Graph Integration: Track freshness history for intelligent recommendations

AI-Powered Semantic Analysis (v0.6.0+)

  • πŸ€– LLM Integration: Optional integration with DeepSeek, OpenAI, Anthropic, or Ollama

  • πŸ” Semantic Code Analysis: Detect behavioral changes beyond syntax using AI

  • πŸ§ͺ Example Validation: Simulate code execution to verify documentation examples

  • 🎯 Intelligent Fallback: Graceful degradation to AST-only analysis when LLM unavailable

  • πŸ”’ Privacy First: Works fully offline with AST analysis, LLM completely optional

Requirements

  • Node.js: 20.0.0 or higher

  • npm: Latest stable version

Installation

Option A β€” npm global install (recommended):

npm install -g documcp

Option B β€” build from source (contributors / local dev):

git clone https://github.com/tosin2013/documcp.git
cd documcp
npm install
npm run build

MCP Client Setup

DocuMCP works with any MCP-enabled AI client. Pick the one you use:

Claude Desktop

File location:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

  • Windows: %APPDATA%\Claude\claude_desktop_config.json

  • Linux: ~/.config/claude/claude_desktop_config.json

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

Restart Claude Desktop to load the configuration.

Claude Code (CLI)

# npm global install
claude mcp add documcp -- npx documcp

# local build (replace path as needed)
claude mcp add documcp -- node /path/to/documcp/dist/index.js

Verify with claude mcp list β€” no restart required.

Cursor

File location: ~/.cursor/mcp.json (or Settings > MCP in the UI)

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

VS Code (GitHub Copilot)

Add to settings.json (Cmd/Ctrl + Shift + P β†’ Preferences: Open User Settings (JSON)):

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

Troubleshooting

  • Ensure npx is available in your PATH

  • To find a global install path: npm list -g documcp

  • To use the full path instead of npx:

    {
      "command": "node",
      "args": ["/usr/local/lib/node_modules/documcp/dist/index.js"]
    }

Quick Start

Once connected to your AI client, use natural language prompts:

Analyze my repository at /path/to/my-project
Recommend a static site generator based on my analysis
Set up a Diataxis documentation structure using Docusaurus
Deploy my documentation to GitHub Pages

DocuMCP provides 30+ tools including repository analysis, intelligent SSG recommendations, content generation, deployment automation with tracking, validation, user preference management, deployment analytics, and memory-enhanced insights. See the complete documentation for detailed tool reference.

Key Tools

Analysis & Recommendations

  • analyze_repository - Deep repository structure and dependency analysis

  • recommend_ssg - Intelligent SSG recommendations with historical data and user preferences

  • detect_gaps - Identify missing documentation sections

Deployment & Tracking

  • deploy_site - Automated deployment to GitHub Pages or Vercel with outcome tracking (deploy_pages still works as a back-compat alias)

  • verify_deployment - Comprehensive deployment validation

  • analyze_deployments - Analytics and insights from deployment history

User Preferences & Learning

  • manage_preferences - Manage user preferences for personalized recommendations

  • View historical success rates and deployment patterns

  • Get recommendations based on similar projects' success

Development

# Run in development mode
npm run dev

# Run tests
npm test

# Lint code
npm run lint

# Type check
npm run typecheck

Architecture

DocuMCP follows a modular, stateless architecture:

  • TypeScript-based implementation using the official MCP SDK

  • Stateless operation for consistency and reliability

  • Modular design with clear separation of concerns

  • Progressive complexity allowing users to start simple

Documentation Structure (Diataxis)

DocuMCP automatically creates documentation following the Diataxis framework:

  • Tutorials: Learning-oriented guides for newcomers

  • How-To Guides: Task-oriented recipes for specific goals

  • Reference: Information-oriented technical descriptions

  • Explanation: Understanding-oriented conceptual discussions

Contributing

We welcome contributions! Please see our Contributing Guide for details.

First Time Contributors

Look for issues labeled "good first issue" to get started with the project. We welcome contributions from developers of all experience levels.

Reporting Issues

Please use our issue templates when reporting bugs or requesting features.

Code of Conduct

This project adheres to the Contributor Covenant Code of Conduct. By participating, you are expected to uphold this code.

Security

Please see our Security Policy for reporting vulnerabilities and security-related issues.

License

MIT License - see LICENSE for details.

Acknowledgments

Available Tools

52 tools
analyze_deploymentsC

Analyze deployment patterns and generate insights from historical deployment data

Inputs: analysisType, ssg, ssgs, periodDays.

analyze_readmeB

Comprehensive README analysis with length assessment, structure evaluation, and optimization opportunities

Inputs: project_path (required), target_audience, optimization_level, max_length_target.

analyze_repositoryC

Analyze repository structure, dependencies, and documentation needs

Inputs: path (required), depth.

change_watcherB

Watch code changes and trigger documentation drift detection in near real-time.

Inputs: projectPath (required), docsPath (required), action, watchPaths, +10 more.

cleanup_agent_artifactsA

Detect, classify, and clean up artifacts generated by AI coding agents (e.g., TODO.md, PLAN.md, agent markers, temporary files). Supports scan, clean, and archive operations with configurable patterns.

Inputs: path (required), operation (required), dryRun, interactive, +3 more.

deploy_pagesB

[DEPRECATED β€” use deploy_site instead] Set up GitHub Pages deployment workflow. This alias will be removed in v1.1.0.

Inputs: repository (required), ssg (required), branch, customDomain, +4 more.

deploy_siteA

Set up deployment workflow for GitHub Pages or Vercel with deployment tracking and preference learning. Use the 'target' parameter to choose one deployment platform per project (github-pages or vercel). If you switch targets, the tool will warn you to remove the old workflow file.

Inputs: repository (required), ssg, target, branch, +6 more.

detect_documentation_gapsC

Analyze repository and existing documentation to identify missing content and gaps

Inputs: repositoryPath (required), documentationPath, analysisId, depth.

evaluate_readme_healthB

Evaluate README files for community health, accessibility, and onboarding effectiveness

Inputs: readme_path (required), project_type, repository_path.

generate_configC

Generate configuration files for the selected static site generator

Inputs: ssg (required), projectName (required), outputPath (required), projectDescription.

generate_contextual_contentC

Generate context-aware documentation using AST analysis and knowledge graph insights (Phase 3)

Inputs: filePath (required), documentationType, includeExamples, style, +1 more.

generate_llm_contextC

Generate a comprehensive LLM context reference file documenting all tools, memory system, and workflows for easy @ reference

Inputs: projectPath (required), includeExamples, format.

generate_readme_templateB

Generate standardized README templates for different project types with best practices

Inputs: projectName (required), description (required), templateType (required), author, +5 more.

get_community_insightsA

Aggregate cross-project signals from the Knowledge Graph and surface community-level insights: deployment success rates per SSG, common technology stacks, frequent drift sources, and project health distribution. All data is anonymized β€” no raw project paths are exposed.

Inputs: no named parameters.

manage_preferencesC

Manage user preferences for documentation generation and SSG recommendations

Inputs: action (required), userId, preferences, json.

manage_sitemapB

Generate, validate, and manage sitemap.xml as the source of truth for documentation links. Sitemap.xml is used for SEO, search engine submission, and deployment tracking.

Inputs: action (required), docsPath (required), baseUrl, includePatterns, +4 more.

memory_agent_networkC

Manage multi-agent memory sharing and collaboration

Inputs: action (required), agentInfo, memoryId, targetAgent, +1 more.

memory_cleanupC

Clean up old memories

Inputs: daysToKeep, dryRun.

memory_enhanced_recommendationC

Get enhanced recommendations using learning and knowledge graph

Inputs: projectPath (required), baseRecommendation (required), projectFeatures (required).

memory_exportC

Export memories to JSON or CSV

Inputs: format, filter.

memory_export_advancedC

Advanced memory export with multiple formats and options

Inputs: outputPath (required), options.

memory_import_advancedC

Advanced memory import with validation and conflict resolution

Inputs: inputPath (required), options.

memory_insightsC

Get insights and patterns from memory

Inputs: projectId, timeRange.

memory_intelligent_analysisC

Get intelligent analysis with patterns, predictions, and recommendations

Inputs: projectPath (required), baseAnalysis (required).

memory_knowledge_graphC

Query the knowledge graph for relationships and paths

Inputs: query (required).

memory_learning_statsC

Get comprehensive learning and knowledge graph statistics

Inputs: includeDetails.

memory_migrationC

Create and execute migration plans between different memory systems

Inputs: action, sourcePath, migrationPlan, sourceSchema, +2 more.

memory_optimization_metricsC

Get comprehensive optimization metrics and recommendations

Inputs: includeRecommendations, timeRange.

memory_pruningC

Intelligent memory pruning and optimization

Inputs: policy, dryRun.

memory_recallC

Recall memories about a project or topic

Inputs: query (required), type, limit.

memory_similarB

Find similar projects from memory

Inputs: analysisId (required), limit.

memory_temporal_analysisD

Analyze temporal patterns and trends in memory data

Inputs: query, analysisType.

memory_visualizationC

Generate visual representations of memory data

Inputs: visualizationType, options, customVisualization.

optimize_readmeB

Optimize README content by restructuring, condensing, and extracting detailed sections to separate documentation

Inputs: readme_path (required), strategy, max_length, include_tldr, +3 more.

populate_diataxis_contentC

Intelligently populate Diataxis documentation with project-specific content

Inputs: analysisId (required), docsPath (required), populationLevel, includeProjectSpecific, +2 more.

read_directoryA

List files and directories within allowed roots. Use this to discover files without requiring full absolute paths from the user.

Inputs: path (required).

readme_best_practicesA

Analyze README files against best practices checklist and generate templates for improvement

Inputs: readme_path (required), project_type, generate_template, output_directory, +2 more.

recommend_ssgA

Recommend the best static site generator based on project analysis and user preferences

Inputs: analysisId (required), userId, preferences.

record_drift_outcomeA

Record the host's judgment ('actionable' | 'noise' | 'deferred') for a previously-detected drift. Outcomes feed back into priority-scoring weights via the knowledge graph (ADR-012 Phase 4, Issue #114).

Inputs: projectPath (required), driftId (required), outcome (required), notes.

setup_playwright_testsC

Generate Playwright E2E test setup for documentation site (containers + CI/CD)

Inputs: repositoryPath (required), ssg (required), projectName (required), mainBranch, +3 more.

setup_structureC

Create Diataxis-compliant documentation structure

Inputs: path (required), ssg (required), includeExamples.

sync_code_to_docsC

Automatically synchronize documentation with code changes using AST-based drift detection (Phase 3)

Inputs: projectPath (required), docsPath (required), mode, autoApplyThreshold, +1 more.

test_local_deploymentB

Test documentation build and local server before deploying to GitHub Pages

Inputs: repositoryPath (required), ssg (required), port, timeout, +1 more.

track_documentation_freshnessB

Scan documentation directory for staleness markers and identify files needing updates based on configurable time thresholds (minutes, hours, days)

Inputs: docsPath (required), projectPath, warningThreshold, staleThreshold, +5 more.

update_existing_documentationC

Intelligently analyze and update existing documentation using memory insights and code comparison

Inputs: analysisId (required), docsPath (required), compareMode, updateStrategy, +2 more.

validate_contentB

Validate general content quality: broken links, code syntax, references, and basic accuracy

Inputs: contentPath (required), validationType, includeCodeValidation, followExternalLinks.

validate_diataxis_contentB

Validate the accuracy, completeness, and compliance of generated Diataxis documentation

Inputs: contentPath (required), analysisId, validationType, includeCodeValidation, +1 more.

validate_documentation_freshnessC

Validate documentation freshness, initialize metadata for files without it, and update timestamps based on code changes

Inputs: docsPath (required), projectPath (required), initializeMissing, updateExisting, +2 more.

validate_readme_checklistB

Validate README files against community best practices checklist with detailed scoring

Inputs: readmePath (required), projectPath, strict, outputFormat.

verify_deploymentB

Verify and troubleshoot GitHub Pages deployment

Inputs: repository (required), url.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    C2.4/5.0

    Scored across 52 tools

    Disambiguation2/5

    Many memory_* tools have overlapping purposes (e.g., memory_recall, memory_contextual_search, memory_similar, memory_insights all retrieve/analyze memories), and there are multiple validation tools (validate_diataxis_content, validate_content, validate_readme_checklist) that could cause misselection. The boundaries between these tools are not clearly distinct, leading to ambiguity.

    Naming Consistency2/5

    The majority follow verb_noun (e.g., analyze_repository, generate_config), but the memory_* group is inconsistentβ€”some are verb-noun (memory_recall), others are adjective-noun (memory_similar), and there's a deprecated alias (deploy_pages) alongside deploy_site. Mixed conventions and occasional non-verb patterns reduce predictability.

    Tool Count1/5

    52 tools is excessive for a documentation MCP server. The core workflow (analysis, SSG, generation, deployment, validation) could be covered with 10-15 tools; the large memory_* cluster and multiple overlapping validation tools inflate the surface and make it unwieldy for an agent to navigate.

    Completeness4/5

    The server covers a broad range: repository analysis, SSG recommendation, configuration, structure setup, content generation, validation, deployment, README analysis, memory management, and drift detection. While some operations like deleting documentation are absent, the coverage is extensive and few critical gaps are apparent for the stated purpose.

    Maintenance

    ActivityNo data
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Transforms any GitHub repository or GitHub Pages site into a documentation hub for AI assistants using the Model Context Protocol. It allows AI tools to access real-time code and documentation to prevent hallucinations and provide accurate API usage examples.
      Apache 2.0
    • A
      license
      Not graded
      quality
      D
      maintenance
      A project-agnostic MCP server that exposes Markdown documentation from a project's /docs folder as MCP resources for AI agents. It provides stable, up-to-date context to reduce hallucinations and ensure agents remain aligned with project-specific conventions and goals.
      MIT
    • A
      license
      Not graded
      quality
      C
      maintenance
      MCP server that fetches documentation from public git repositories at specific versions, enabling AI coding assistants to access up-to-date, ref-pinned docs from source repos.
      2
      Apache 2.0
    • A
      license
      A
      quality
      D
      maintenance
      Comprehensive documentation lifecycle management powered by an MCP server. Automates documentation creation, maintenance, quality assessment, and synchronization for software projects.
      8
      6
      MIT