A practical baseline for building retrieval-augmented generation features without copying vendor documentation wholesale.
Intended use
Use this Block when you need chunking, embedding, hybrid search, and grounded answer assembly in a greenfield AI feature.
What changed in 1.1.0
- Documented hybrid lexical + semantic fusion and when to add reranking.
- Expanded verification steps for export and search evaluation.
- Clarified trust boundaries for retrieved content.