Welcome to Fairness Scanner — a full-stack AI governance interface for identifying and mitigating fairness risks in machine learning and LLM systems.

This project includes:

• ✅ A modern Next.js frontend for dataset upload and visual analysis
• ✅ A lightweight FastAPI backend powered by Giskard for automated fairness evaluation
• ✅ Optional support for MinIO, Spark, and Dremio for scalable data lake integration

• 📂 Upload CSV or Excel datasets via the web UI
• 🔍 Automatically detect protected attributes (e.g., gender, age, race)
• ⚖️ Run fairness diagnostics using Giskard
• 📊 Visualize results and trigger mitigation workflows
• ☁️ Optional support for S3-compatible object storage (MinIO)
• 🔌 Extendable to integrate with Flink, Spark, Dremio, and LLM agents

├── frontend/         # Next.js (React) frontend
│   ├── app/
│   ├── lib/
│   ├── components/
│   ├── public/
│   └── ...
├── backend/          # FastAPI + Giskard backend
│   ├── main.py
│   ├── upload.py
│   └── requirements.txt
├── setup.sh          # Dev environment setup script
└── README.md

Specify the backend endpoint for the frontend using NEXT_PUBLIC_API_URL in frontend/.env.local:

NEXT_PUBLIC_API_URL=http://localhost:8000

This value is read by the browser when uploading datasets.