Enterprise RAG Studio is a production-grade, fully observable Retrieval-Augmented Generation platform. Every stage of the RAG pipeline — parsing, chunking, embedding, retrieval, reranking, prompting, generation, and evaluation — is inspectable, measurable, and reproducible.
See docs/00-index.md for the full documentation set and required reading order before contributing.
backend/ FastAPI application (API, services, repositories, AI engines)
frontend/ Next.js application
worker/ Celery worker processes (document, chunk, embedding, evaluation)
docker/ Dockerfiles and compose configuration
docs/ Architecture, database, API, task, and rule documentation
scripts/ Developer and operational scripts
tests/ Cross-service integration/e2e tests
Prerequisites: Docker and Docker Compose.
cp .env.example .env
docker compose -f docker/docker-compose.yml up --buildServices:
| Service | URL |
|---|---|
| Frontend | http://localhost:3000 |
| Backend | http://localhost:8000 |
| API Docs | http://localhost:8000/docs |
| MinIO Console | http://localhost:9001 |
Start at docs/00-index.md. Documentation must be read in the order it defines before making changes.
See CONTRIBUTING.md.
See LICENSE.