dcisionai-mcp-server 2.5.0

DcisionAI MCP Server v2.5 - Enhanced with KB Integration, Retry Mechanism, Validation Feedback, and 3-Industry Wedge Focus (Manufacturing, Retail, Finance)

DcisionAI MCP Server

🚀 AI-Powered Mathematical Optimization for Cursor IDE

The DcisionAI MCP Server brings advanced mathematical optimization capabilities directly to your Cursor IDE. Transform natural language problem descriptions into optimal solutions using state-of-the-art AI models and robust optimization solvers.

✨ Features

• 8 Powerful Tools: Complete optimization workflow from problem understanding to business explanations
• AI-Driven Problem Formulation: Uses Claude 3 Haiku to translate business problems into mathematical models
• Real Optimization Solvers: OR-Tools integration with PDLP, GLOP, CBC, SCIP, and more
• Business Explainability: Generate executive summaries and implementation guidance
• 21 Industry Workflows: Pre-built templates for manufacturing, healthcare, finance, and more
• Cursor IDE Integration: Seamless integration with Cursor's MCP protocol

🛠️ Available Tools

1. classify_intent - Understand and classify optimization problems
2. analyze_data - Assess data quality and identify variables/constraints
3. build_model - Generate mathematical optimization models using AI
4. select_solver - Choose the best solver for your problem type
5. solve_optimization - Execute optimization using real solvers
6. explain_optimization - Generate business-friendly explanations
7. get_workflow_templates - Access 21 industry-specific workflows
8. execute_workflow - Run complete optimization workflows

🚀 Quick Start

Installation

# Install via pip
pip install dcisionai-mcp-server

# Or use uvx for direct execution
uvx dcisionai-mcp-server@latest

Cursor IDE Setup

Add to your ~/.cursor/mcp.json:

{
  "mcpServers": {
    "dcisionai-mcp-server": {
      "command": "uvx",
      "args": ["dcisionai-mcp-server@latest"],
      "env": {
        "PYTHONUNBUFFERED": "1"
      },
      "disabled": false,
      "autoApprove": [
        "classify_intent",
        "analyze_data", 
        "build_model",
        "solve_optimization",
        "select_solver",
        "explain_optimization",
        "get_workflow_templates",
        "execute_workflow"
      ]
    }
  }
}

Usage Example

# In Cursor IDE, use the MCP tools:
@dcisionai-mcp-server classify_intent "Optimize my investment portfolio for maximum returns with moderate risk"

# Follow up with:
@dcisionai-mcp-server build_model "Portfolio optimization problem" --intent_data <previous_result>

# Continue the workflow:
@dcisionai-mcp-server solve_optimization "Portfolio problem" --model_building <model_result>

📊 Supported Optimization Types

• Linear Programming (LP) - Resource allocation, production planning
• Mixed-Integer Linear Programming (MILP) - Scheduling, routing
• Quadratic Programming (QP) - Portfolio optimization, risk management
• Convex Optimization - Machine learning, signal processing

🏭 Industry Workflows

• Manufacturing: Production planning, inventory optimization, quality control
• Healthcare: Staff scheduling, patient flow, resource allocation
• Finance: Portfolio optimization, risk assessment, fraud detection
• Retail: Demand forecasting, pricing optimization, supply chain
• Logistics: Route optimization, warehouse management, fleet operations
• Energy: Grid optimization, renewable integration, demand response
• Marketing: Campaign optimization, budget allocation, customer segmentation

🔧 Requirements

• Python 3.8+ (Python 3.13 has limited OR-Tools support)
• AWS credentials for Bedrock access (for AI model inference)
• Cursor IDE (for MCP integration)

📚 Documentation

• Platform Overview
• API Reference
• Quick Start Guide
• Deployment Guide

🤝 Contributing

We welcome contributions! Please see our Contributing Guidelines.

📄 License

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

🆘 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Email: [e-mail]

🙏 Acknowledgments

• OR-Tools for optimization solvers
• Claude 3 Haiku for AI model inference
• Cursor IDE for MCP protocol support
• AWS Bedrock for AI model hosting

---

Made with ❤️ by the DcisionAI Team