Repo · Search & RAG

jev-corrective-rag

Corrective RAG demo where triage, per-chunk grading, and answer verification use TypeSafe System One (Jev) typed gates instead of LLM judges;…

Open github.com ↗

How builders describe it

Corrective RAG demo where triage, per-chunk grading, and answer verification use TypeSafe System One (Jev) typed gates instead of LLM judges; generation stays on a separate LLM. Includes a Streamlit UI, CLI, and bench harness with keyword-stub fallback when no TypeSafe key is set.

The decision Jev makes

Rank candidates against a query, or answer whether a passage is relevant.

Where it fits

Search pipelines retrieve too much. These projects use Jev as a reranker or relevance check after the first search, so only useful results reach the user or the model. All 56 search & rag projects →

Related projects