Claude Deep Think MCP Server
A powerful Model Context Protocol (MCP) server that provides proactive deep analytical thinking using Anthropic's Claude Sonnet 4.5. This tool is designed to be called BEFORE writing code when new information arrives.
ā” Core Concept: Think Before Code
Use this tool FIRST when new information arrives, BEFORE writing any code:
š Error messages or stack traces
š User requirements or feature requests
š¬ Code review feedback
š Performance issues
š Security alerts
š API documentation to integrate
šļø Database problems
š UX/UI feedback
š Breaking changes in dependencies
šļø Architectural decisions
Workflow: New Info ā Think Tool ā Review Insights ā Write Better Code
š Features
Deep Think & Analysis (claude_think)
Provides intelligent insights, suggestions, and strategic guidance before code implementation. Perfect for:
ā Understanding context deeply before acting
ā Identifying potential pitfalls upfront
ā Suggesting best practices from the start
ā Offering alternative approaches
ā Extracting key information for efficient implementation
ā Strategic decision-making
ā Problem-solving and architecture planning
Result: Fewer bugs, better code quality, faster development!
š Prerequisites
Node.js 18+ or Bun
Anthropic Claude API key (Get one here)
MCP-compatible client (Cursor IDE, Claude Desktop, etc.)
š Quick Start
1. Installation
2. Configuration
The API key is configured when connecting to the MCP server (see Docker or Cursor setup below).
3. Build
š³ Docker Setup (Recommended)
Quick Start
The container will auto-restart when Docker Desktop launches.
Docker Configuration
The server runs on http://localhost:8080/mcp with the following environment variables:
ANTHROPIC_API_KEY- Your Claude API key (required)CLAUDE_MODEL- Model to use (default: claude-sonnet-4-20250514)
š§ Usage in Cursor IDE
Docker Connection (Recommended)
Add to your ~/.cursor/mcp.json or .cursor/mcp.json:
Enable Proactive Thinking
Copy the .cursorrules file from this repo to your project root. This makes Cursor AI automatically use the think tool before writing code.
Tool Usage Pattern
Always use this pattern when new information arrives:
Examples:
Error Message:
New Feature:
Performance Issue:
š Examples
Example 1: Analyzing Technical Decisions
Expected Output: Comprehensive comparison with pros, cons, and recommendations
Example 2: Architecture Planning
Example 3: Best Practices
š ļø Development
Project Structure
Available Scripts
bun run build/npm run build- Compile TypeScriptbun run dev/npm run dev- Development server with hot reload
š Security Best Practices
Never commit API keys - Always use environment variables
Use .gitignore - Ensure
.envfiles are ignoredRotate keys regularly - Update API keys periodically
Review tool calls - Keep manual approval enabled in Cursor
Use development environments - Test with non-production data
š¦ Docker Management
š Troubleshooting
Issue: Server not connecting in Cursor
Solutions:
Verify Docker container is running:
docker ps | grep claude-visionCheck container logs:
docker logs claude-vision-mcp-serverRestart Cursor IDE completely
Verify API key in URL is correct
Issue: API key errors
Solutions:
Ensure key starts with
sk-ant-Test key at: https://console.anthropic.com/
Check environment variables in container
Verify URL parameter format
Issue: Container won't start
Solutions:
š” Performance
With Bun runtime:
ā” 4x faster package installs
ā” 3-4x faster script execution
š¦ Smaller Docker images
š Faster cold starts
š Comprehensive Guides
PROACTIVE_THINKING_WORKFLOW.md - Complete workflow guide with before/after examples
THINK_TOOL_EXAMPLES.md - 10 real-world usage examples
.cursorrules - Cursor IDE rules for automatic think-before-code pattern
š” Why This Workflow?
Without Think Tool:
With Claude_Think Tool:
Result: 6x faster, better quality, fewer bugs! š
š License
MIT
š¤ Contributing
Contributions are welcome! Please feel free to submit a Pull Request.
š Support
For issues or questions:
Open an issue on GitHub
Check the MCP Documentation
Read the workflow guides in this repository
š Acknowledgments
Built with Anthropic Claude API
Powered by Model Context Protocol
Containerized with Bun
This server cannot be installed