fairmind

FairMind is an open‑source ethical AI governance platform that aims to provide evidence‑grade evaluation for machine learning, large language models, and multimodal AI systems. The project is currently in an internal alpha stage, offering a foundational control plane for immutable target and suite identity, execution‑envelope binding, and trust‑policy verification. It includes legacy bias detection endpoints for predictive models (demographic parity, equalized odds, disparate impact, individual and group fairness) and planned LLM and multimodal bias lanes (WEAT, SEAT, minimal pairs, image/audio/video generation bias). The platform also provides remediation code generation, optional MLOps integrations with Weights & Biases and MLflow, governance evidence mapping for frameworks such as EU AI Act and ISO/IEC 42001, a model registry with versioning, and monitoring dashboards. The backend is built with FastAPI, Python, SQLAlchemy, SQLite/DuckDB/PostgreSQL, Redis, and JWT authentication, while the frontend uses Next.js, React, TypeScript, and a neobrutalist UI. Development instructions, API documentation, and a detailed project structure are included in the repository.