MCP Prompt Optimizer

A professional-grade MCP (Model Context Protocol) server that provides cutting-edge prompt optimization tools with research-backed strategies delivering 15-74% performance improvements.

- Clarity: Simplifies prompts for directness and precision
- Specificity: Adds detailed constraints and requirements
- Chain of Thought: Incorporates step-by-step reasoning
- Few-Shot: Includes example formats for guidance
- Structured Output: Defines clear output organization
- Role-Based: Adds expert role context

- Tree of Thoughts (ToT): Multi-path reasoning with 74% success rate on complex tasks
- Constitutional AI: Self-critique and alignment with safety principles
- Automatic Prompt Engineer (APE): AI-discovered optimal instruction patterns
- Meta-Prompting: AI generates its own optimized prompts
- Self-Refine: Iterative improvement with 20% performance gains
- TEXTGRAD: Natural language feedback as optimization gradients
- Medprompt: Multi-technique ensemble achieving 90%+ accuracy
- PromptWizard: Feedback-driven self-evolving prompts

Production-ready templates across 11 domains:

- Business Analysis: Competitive analysis frameworks
- Product Management: User research synthesis
- Content Creation: Technical blog posts with SEO optimization
- Development: Comprehensive code review checklists
- Communication: Stakeholder updates and project reports
- Strategy: OKR planning frameworks
- Operations: Standard Operating Procedures (SOPs)
- Legal: Contract termination and compliance
- Customer Experience: Feedback surveys and insights
- Data Analysis: Data insights and reporting
- Meeting Management: Effective meeting agendas

# Clone the repository git clone <repository-url> cd mcp-prompt-optimizer # Create virtual environment (recommended) python3 -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate # Install dependencies ./install.sh # Or install manually pip install -r requirements.txt # Configure Claude Desktop python3 setup_interactive.py

Add to your Claude Desktop configuration file:

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

Windows:%APPDATA%\Claude\claude_desktop_config.json

Linux:~/.config/Claude/claude_desktop_config.json

{ "mcpServers": { "prompt-optimizer": { "command": "python3", "args": ["/path/to/mcp-prompt-optimizer/prompt_optimizer.py"], "env": {} } } }

# Analyze prompt quality "Analyze this prompt: write a blog post about AI" # Apply specific optimization "Optimize this prompt using chain_of_thought: explain machine learning" # Auto-select best strategy "Auto-optimize: help me debug this code" # Get domain template "Get domain template for code_review_checklist"

# Use Tree of Thoughts for complex problems "Apply advanced optimization with tree_of_thoughts: design a microservices architecture" # Use Constitutional AI for safety-critical tasks "Apply advanced optimization with constitutional_ai: create content moderation guidelines" # Use Medprompt for high-accuracy classification "Apply advanced optimization with medprompt: categorize customer support tickets" # List available templates "List all domain templates"

mcp-prompt-optimizer/ ├── prompt_optimizer.py # Main MCP server ├── advanced_strategies.py # Research-backed optimization strategies ├── domain_templates.py # Professional domain templates ├── examples.py # Usage examples and demonstrations ├── setup_interactive.py # Automated setup script └── README.md # This file

# Run basic tests ./test.sh # Run usage examples python3 examples.py

- analyze_prompt: Analyzes prompt quality and identifies issues
- optimize_prompt: Applies specific optimization strategies
- auto_optimize: Automatically selects optimal strategy
- get_prompt_template: Returns basic templates
- advanced_optimize: Applies research-backed strategies
- get_domain_template: Returns professional domain templates
- list_domain_templates: Lists available templates by domain
- Fork the repository
- Create a feature branch
- Add tests for new functionality
- Update documentation
- Submit a pull request

- New Strategy: Add toadvanced_strategies.py
- New Template: Add todomain_templates.py
- Examples: Add toexamples.py

- Check Python version:python3 --version(requires 3.8+)
- Install dependencies: Run./install.shorpip install -r requirements.txt
- Verify MCP installation:pip show mcp
- Check Claude Desktop logs
- Restart Claude Desktop

- Verify configuration file location
- Check file paths in configuration
- Run setup script again

# Test server directly python3 prompt_optimizer.py # Verbose logging export MCP_LOG_LEVEL=debug python3 prompt_optimizer.py

This project is licensed under the MIT License - see theLICENSEfile for details.

@software{mcp_prompt_optimizer, title={MCP Prompt Optimizer: Research-Backed Prompt Optimization for AI Systems}, author={Bubobot}, year={2024}, url={https://github.com/Bubobot-Team/mcp-prompt-optimizer} }

For questions, issues, or contributions, please visit ourGitHub repository.

A TypeScript agent that integrates MCP servers with Ollama, allowing AI models to use various tools through a unified interface.