Ultimate MCP Server v2.0

The definitive all-in-one Model Context Protocol (MCP) server for AI-assisted coding across 30+ platforms.

🚀 Features

Core Capabilities

• 🤖 50+ AI Models: Latest models including GPT-4o, Claude 3 Opus, Gemini 2.5, DeepSeek V3, Grok-4, and more
• 🔌 Multi-Transport Support: STDIO, SSE, HTTP/REST, WebSocket - works everywhere
• 📚 RAG System: Document-based AI with vector search and embeddings
• 🧠 Cognitive Memory: Knowledge graphs with persistent context
• 💻 Code Intelligence: AST parsing, symbol extraction, dependency analysis
• 🔍 Universal Search: Cross-platform file and code search
• 📊 Advanced Analytics: Performance monitoring and cost optimization
• 🌐 Browser Automation: Playwright and Puppeteer integration
• 🎨 UI Understanding: Visual analysis and design system extraction

Platform Compatibility

✅ Fully Tested: Claude Desktop, Claude Code, Cursor, VS Code (Continue), Cline, Windsurf, Google AI Studio

🧪 In Testing: Smithery, Trae, Visual Studio 2022, Zed, BoltAI, Augment Code, Roo Code, and 20+ more

📦 Installation

Quick Start (Recommended)

npx ultimate-mcp-server

Add to Claude Code

claude mcp add npx ultimate-mcp-server

Global Installation

npm install -g ultimate-mcp-server
ultimate-mcp-server

Local Installation

npm install ultimate-mcp-server

🔧 Configuration

Claude Desktop

Add to your Claude Desktop config file:

{
  "mcpServers": {
    "ultimate-mcp": {
      "command": "npx",
      "args": ["ultimate-mcp-server"]
    }
  }
}

Cursor/Windsurf

Add to .cursorrules or .windsurfrules:

{
  "mcpServers": {
    "ultimate-mcp": {
      "command": "npx",
      "args": ["ultimate-mcp-server"]
    }
  }
}

Environment Variables

Create a .env file with your API keys:

# Required (at least one)
OPENAI_API_KEY=your-key
ANTHROPIC_API_KEY=your-key
GOOGLE_API_KEY=your-key

# Optional (for specific features)
PERPLEXITY_API_KEY=your-key  # For research features
XAI_API_KEY=your-key         # For Grok models
MISTRAL_API_KEY=your-key     # For Mistral models

🛠️ Available Tools

Debugging & Analysis

• analyze_error - Analyze errors with AI-powered suggestions
• explain_code - Get detailed code explanations
• suggest_optimizations - Performance and code quality improvements
• debugging_session - Interactive debugging with context

AI Orchestration

• ask - Query any AI model directly
• orchestrate - Coordinate multiple models for complex tasks
• compare_models - Compare responses across different models

Code Generation

• generate_code - Create code with best practices
• analyze_codebase - Large-scale codebase analysis
• find_in_codebase - Pattern search across files

Advanced Features

• rag_search - Semantic search in documents
• build_knowledge_graph - Create knowledge representations
• analyze_ui_screenshot - Understand UI/UX from images
• capture_webpage_screenshot - Browser automation
• extract_webpage_data - Web scraping

🎯 Use Cases

1. Code Review & Debugging

// Analyze an error
await mcp.use_tool('analyze_error', {
  error: 'TypeError: Cannot read property of undefined',
  code: problemCode,
  language: 'javascript'
});

2. UI/UX Analysis

// Analyze a design screenshot
await mcp.use_tool('analyze_ui_screenshot', {
  url: 'https://example.com',
  analysis_type: 'comprehensive',
  extract_design_system: true
});

3. Large Codebase Analysis

// Analyze entire project
await mcp.use_tool('analyze_large_codebase', {
  rootDir: '/path/to/project',
  pattern: '.*\\.ts$',
  query: 'How is authentication implemented?'
});

4. Multi-Model Orchestration

// Get consensus from multiple models
await mcp.use_tool('orchestrate', {
  prompt: 'Design a scalable microservices architecture',
  strategy: 'consensus',
  models: ['gpt-4o', 'claude-3-opus', 'gemini-2.5-pro']
});

🏗️ Architecture

Smart Lazy Loading

• Tools are registered immediately but loaded on-demand
• Reduces startup time from 5s to <500ms
• Maintains full functionality without "tools not available" issues

Performance Optimization

• Intelligent model routing based on task complexity
• Automatic cost optimization with quality thresholds
• Built-in caching and rate limiting
• Token usage tracking and optimization

Extensibility

• Plugin architecture for custom tools
• Support for custom embedding providers
• Configurable vector stores (Pinecone, Weaviate, ChromaDB)
• Custom model integrations

📊 Performance Metrics

• Startup Time: <500ms (with lazy loading)
• Response Time: <1s for most operations
• Memory Usage: <512MB under normal load
• Bundle Size: ~45MB (optimized)
• Context Window: Up to 2M tokens (model dependent)

🔐 Security

• No data persistence by default
• API keys stored locally only
• Sandboxed code execution
• Rate limiting and abuse prevention
• Audit logging for compliance

🤝 Contributing

We welcome contributions! See CONTRIBUTING.md for guidelines.

📄 License

MIT License - see LICENSE for details.

🙏 Acknowledgments

Ultimate MCP incorporates the best features from:

• agentset-ai - RAG capabilities
• contentful-mcp - Content management
• cognee-mcp - Cognitive memory
• code-context-provider - Code analysis
• code-assistant - Autonomous exploration
• mcp-enhance-prompt - Prompt engineering
• mcp-everything-search - Universal search
• consult7 - Large context analysis

🚀 What's New in v2.0

• 50+ Latest AI Models: Including Grok-4, DeepSeek V3, Gemini 2.5 Flash
• Browser Automation: Playwright and Puppeteer integration
• UI/UX Understanding: Analyze designs from screenshots or URLs
• Smart Lazy Loading: 10x faster startup with full functionality
• Cost Optimization: Automatic model selection based on budget
• Performance Monitoring: Real-time metrics and insights
• Large Context Analysis: Handle massive codebases (Consult7-style)
• Multi-Transport: Works with every MCP-compatible platform

Built with ❤️ by the Ultimate MCP Team