dcisionai-mcp-server

DcisionAI MCP Server

AI-Powered Optimization for Cursor, Claude Desktop & VS Code

Solve complex optimization problems directly in your IDE using natural language. Get mathematically-verified solutions with 90%+ trust scores in seconds.

🚀 Quick Start

Installation (Zero Configuration!)

# That's it! No installation needed with uvx

Configure Your IDE

For Cursor or Claude Desktop:

Add to your MCP config file (~/.cursor/mcp.json on Mac):

{
  "mcpServers": {
    "dcisionai-optimization": {
      "command": "uvx",
      "args": ["dcisionai-mcp-server@latest"],
      "env": {
        "DCISIONAI_API_URL": "https://dcisionai-mcp-platform-production.up.railway.app"
      },
      "autoApprove": ["dcisionai_solve", "dcisionai_solve_with_model"]
    }
  }
}

Use It!

In Cursor or Claude Desktop, just ask:

"Use DcisionAI to optimize my $500K portfolio concentrated in tech stocks"

"Use DcisionAI to optimize delivery routes for 20 customers"

"Use DcisionAI to optimize employee scheduling for 30 workers across 50 shifts"

✨ What Can It Do?

📊 Finance

• Portfolio rebalancing with risk constraints
• Trading schedule optimization
• Asset allocation with concentration limits
• Private equity exit timing

🏪 Retail

• Store layout optimization (shelf space allocation)
• Inventory management
• Pricing optimization
• Supply chain optimization

🏭 Manufacturing

• Production scheduling
• Resource allocation
• Job shop optimization
• Workforce scheduling

🚚 Logistics

• Vehicle routing (VRP)
• Delivery route optimization
• Warehouse layout
• Distribution network design

🛠️ Tools

• dcisionai_solve - Full optimization workflow (classification, intent extraction, solving, explanation)
• dcisionai_solve_with_model - Solve using deployed models (faster for known problem types)

📚 Resources

• dcisionai://models/list - Available deployed models
• dcisionai://solvers/list - Available solvers (HiGHS, SCIP, DAME, OR-Tools)

🔧 Configuration

Set environment variables:

• DCISIONAI_API_URL: Backend API URL (default: http://localhost:8000)
• DCISIONAI_API_KEY: API key for authentication (optional)
• DCISIONAI_DOMAIN_FILTER: Domain filter ("all", "ria", "pe", etc.)
• DCISIONAI_LOG_LEVEL: Logging level ("INFO", "DEBUG", etc.)

📖 Documentation

• Full Documentation
• MCP Server Planning
• Architecture Decision Record

🤝 Contributing

Contributions welcome! See our GitHub repository for details.

📄 License

MIT License - see LICENSE file for details.

🔗 Links

• Homepage: https://dcisionai.com
• Repository: https://github.com/ameydhavle/dcisionai-mcp-platform
• Issues: https://github.com/ameydhavle/dcisionai-mcp-platform/issues