RAG Lab
Learn how apps search and answer from documents.
- Platform
- Python · Local / Codespaces
- Kind
- Learning tools
- Status
- Open source
- Project year
- 2026
Product identity
The idea
A lab instead of another lecture.
RAG Lab teaches retrieval-augmented generation through missions that ask you to debug real systems. Inspect the intermediate state, change the configuration, and see what the automated checks accept.
The learning loop
Failure is the material.
Work through ingestion, chunking, retrieval and evaluation with an inspectable pipeline. The repository describes thirteen missions and a capstone, with offline operation and automated grading.
Availability
Run it from the source.
The public repository contains setup instructions. This is a learning project, not a hosted service or proof of production retrieval quality. MIT licensed.