Use Cases / rag-relevance
Jev RAG Relevance
Retrievers return near-misses. Stuffing them into the LLM wastes tokens and confuses answers.
Try a related template free on this host — 5 anonymous real-API runs with no signup — then ship the same schema with a prepaid jv_live_ key. See is Jev free? and playground.
Why Jev
Score or noul each chunk for relevance/keep-drop so compaction stays typed — not a lossy rewrite of the evidence.
How to define the decision
For each retrieved chunk, state = query + chunk text. Ask a Noul or Score for relevance / keep. Batch chunks when your worker can parallelize decide calls.
When to use Jev
- Post-retrieve keep / truncate / drop
- Multi-corpus routing before generation
- High-volume RAG where most chunks are noise
When not to use Jev
- Generating the final user-facing answer
- Embedding / index construction itself
- Long creative synthesis across many sources
Sample input and output
From the live tool template. Run it in the playground for a real API response — sample labels below are illustrative until you click Run.
Input
query: How do I rotate my API key? passage: To rotate a key, open Settings > API keys, click Revoke, then Create new key.
Sample display fields
- relevant: yes · 95%
- relevance: direct answer · 92%
Drop low-relevance chunks before prompt assembly. Borderline scores → keep a smaller excerpt or send to a cheaper re-ranker. Never invent missing evidence.
Failure modes and human fallback
- Upstream or network errors: retry with backoff; do not invent a decision.
- Low confidence: route to a human queue or safer default — never auto-act on money/destructive tools.
- Schema drift: keep a golden fixture set in CI so label changes fail the build before production.
- Jev does not embed or index documents — only judges relevance of text you pass.
- Final answer generation stays on a generative LLM.
- Fail closed: if decide errors, skip the chunk or fall back to your current top-k policy.
API example
Same prepaid jv_live_ balance as tools. Output tokens are free on this host's metering model.
curl https://jevtypesafe.org/api/v1/decide \
-H "Authorization: Bearer jv_live_YOUR_KEY_HERE" \
-H "Content-Type: application/json" \
-d '{"state":"RAG check\nQuery: How do I rotate my API key?\nPassage:\nTo rotate a key, open Settings > API keys, click Revoke, then Create new key.","questions":{"relevant":{"type":"noul","instructions":"Does this passage help answer the query?"},"relevance":{"type":"choice","instructions":"How does the passage relate to the query?","criteria":{"direct answer":"directly answers","related":"related but incomplete","tangential":"weakly related","irrelevant":"does not help"}},"supports_claim":{"type":"noul","instructions":"Does the passage support answering without inventing facts?"}}}'FAQ
- Is there a one-click tool?
- Open RAG Relevance under Tools when available, or wire relevance questions to /api/v1/decide in your retriever worker.
Related use cases
Ready to build?
Free anonymous tool runs, then prepaid packs and a jv_live_ key — independent host, not affiliated with TypeSafe AI.