OpenXAI MCP Server

# OpenXAI MCP Server - Submission Guide This guide explains how to submit the OpenXAI MCP server to the [cursor.directory/mcp](https://cursor.directory/mcp) directory. ## Server Overview **Name**: OpenXAI MCP Server **Description**: A Model Context Protocol server for OpenXAI, providing comprehensive tools for evaluating and benchmarking post hoc explanation methods in AI models. **Category**: Machine Learning / Explainable AI **Repository**: https://github.com/yourusername/openxai-mcp ## Key Features 🔍 **Explanation Methods** - LIME (Local Interpretable Model-agnostic Explanations) - SHAP (SHapley Additive exPlanations) - Integrated Gradients - Grad-CAM - Guided Backpropagation 📊 **Evaluation Metrics** - Faithfulness metrics (PGI, PGU) - Stability metrics (RIS, RRS, ROS) - Ground truth metrics (FA, RA, SA, SRA, RC, PRA) - Fairness analysis across subgroups 🗂️ **Datasets** - Synthetic datasets with ground truth explanations - Real-world datasets (German Credit, COMPAS, Adult Income) - Support for tabular, image, and text data 🤖 **Pre-trained Models** - Neural Networks, Logistic Regression, Random Forest - Support Vector Machine, XGBoost ## Installation Instructions ### Prerequisites - Node.js 18+ - Python 3.7+ with OpenXAI package ### Quick Setup ```bash # Clone the repository git clone https://github.com/yourusername/openxai-mcp.git cd openxai-mcp # Install dependencies npm install # Install OpenXAI Python package pip install openxai # Test the server npm test ``` ### Cursor Configuration Add to your Cursor settings (`~/.cursor/mcp.json`): ```json { "mcpServers": { "openxai": { "command": "node", "args": ["/path/to/openxai-mcp/index.js"], "env": {} } } } ``` ## Available Tools 1. **list_datasets** - List available datasets in OpenXAI 2. **load_dataset** - Load a specific dataset 3. **list_models** - List available pre-trained models 4. **load_model** - Load a specific model 5. **list_explainers** - List explanation methods 6. **generate_explanation** - Generate model explanations 7. **list_metrics** - List evaluation metrics 8. **evaluate_explanation** - Evaluate explanation quality 9. **get_leaderboard** - Get benchmarking results 10. **get_framework_info** - Get OpenXAI framework information ## Usage Examples ### Basic Usage ``` Tell me about the OpenXAI framework ``` ### Dataset Exploration ``` List all tabular datasets available in OpenXAI ``` ### Model Analysis ``` Load the German Credit dataset and show me available models for it ``` ### Explanation Generation ``` Generate LIME explanations for the Adult Income dataset using an XGBoost model ``` ### Evaluation ``` Show me all faithfulness metrics available for evaluating explanations ``` ### Benchmarking ``` Get the current leaderboard for explanation methods on the COMPAS dataset ``` ## Technical Details - **Framework**: Model Context Protocol (MCP) 0.6.0 - **Language**: JavaScript/Node.js - **Dependencies**: @modelcontextprotocol/sdk, axios, zod - **Backend**: OpenXAI Python framework - **License**: MIT ## Submission Information ### Repository Structure ``` openxai-mcp/ ├── index.js # Main MCP server implementation ├── package.json # Node.js dependencies ├── README.md # Comprehensive documentation ├── LICENSE # MIT license ├── test.js # Test suite └── SUBMISSION_GUIDE.md # This file ``` ### Validation Checklist - ✅ All 10 tools implemented and tested - ✅ Comprehensive documentation with examples - ✅ MIT license included - ✅ Package.json with proper metadata - ✅ Test suite for verification - ✅ Clear installation instructions - ✅ Integration with OpenXAI framework ### Tags for Submission - `explainable-ai` - `xai` - `machine-learning` - `evaluation` - `benchmarking` - `lime` - `shap` - `openxai` - `explanation-methods` - `model-interpretation` ## Benefits for Users 1. **Researchers**: Benchmark explanation methods against established metrics 2. **ML Engineers**: Evaluate model explanations in production systems 3. **Students**: Learn about explainable AI through hands-on examples 4. **Practitioners**: Compare different explanation methods for their use cases ## OpenXAI Integration This MCP server is built on top of the OpenXAI framework: - **Website**: https://open-xai.github.io/ - **GitHub**: https://github.com/AI4LIFE-GROUP/OpenXAI - **Paper**: https://arxiv.org/abs/2206.11104 ## Support For issues and questions: - GitHub Issues: https://github.com/yourusername/openxai-mcp/issues - OpenXAI Documentation: https://open-xai.github.io/ - OpenXAI Team: openxaibench@gmail.com ## Submission Steps 1. **Create GitHub Repository** - Upload all files to a public GitHub repository - Ensure README.md is comprehensive - Add proper tags and description 2. **Test Locally** - Run `npm test` to verify all tools work - Test with Cursor to ensure MCP integration works - Verify Python OpenXAI package integration 3. **Submit to cursor.directory/mcp** - Visit https://cursor.directory/mcp - Click "Submit MCP Server" - Provide repository URL and description - Include all relevant tags 4. **Follow Up** - Monitor for approval/feedback - Address any issues or suggestions - Update documentation as needed This OpenXAI MCP server provides a valuable bridge between the powerful OpenXAI framework and the MCP ecosystem, enabling easy access to state-of-the-art explainable AI tools through a standard interface.