Use cases· Last updated 2026-09-21

RAG decision workflow with Jev

Retrievers hope. After retrieval, Jev marks relevance, contradiction, or injection; code keeps, flags, or drops passages before a generator sees them.

This unofficial page is the decision workflow slice of the RAG passage decisions pack. Intent: apply the Jev (TypeSafe System One) decision model to RAG passage decisions decision workflow. Primary search language: RAG Jev decision workflow. Confirm patterns on docs.typesafe.ai. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.

Independent angle (cover ≠ clone): Filter/rerank recipes with failure modes; cite vs generate boundary. We cover the intent, not a rival rerank-passages-score URL tree.

RAG use-case context

RAG passage decisions is a workflow, not a chat. Assemble a narrow state, ask the primitives below, and let the retrieval filter branch. TypeSafe’s docs say a good question is a snap decision a knowledgeable person could make in a few seconds — not an open-ended analysis of the query + retrieved passage.

Hub: Classifying RAG passages. Compare, when the other tool is the real job: RAG pipelines.

Decision Workflow inputs

Keep only fields the questions name:

{
  "query": "What is the refund window for pro plans?",
  "passage": { "id": "doc-88#p3", "text": "Pro subscribers may request a refund within 14 days." },
  "corpus": { "trust": "internal_kb" }
}

Point instructions at query, passage.text. Drop the entire 40-passage dump in one state (filter first or ask per pair); embeddings or vector ids Jev cannot use.

Decision signals and actions

All of these share state and run in parallel. Code owns the graph:

def keep(ans, passage_id):
    if ans["injection"].noul >= T_INJECT:
        return "drop_security"
    if ans["relevant"].noul < T_REL:
        return "drop_irrelevant"
    return "keep"

Do not treat a Noul of 0.5 as a “medium” RAG passage decisions score — it means yes and no are equally likely. Conjunctions stay in your code.

Guardrails and escalation

TypeSafe’s confidence-gated examples use a lower bar for recoverable reads than for irreversible actions. Those numbers are illustrations. For RAG passage decisions, treat feed_to_answerer as the high bar (showing a passage to a customer-facing answerer). Tune on labels — see offline evaluation.

Low confidence, or a policy miss → human or safe default; do not feed the passage to the customer-facing answerer.

Evaluation and rollout notes

Shadow: Keep current retriever+LLM; log Jev keep/drop.

Canary: Enforce drops on injection only; relevance stays advisory.

Pin jev-1.13.0 (the versioned id) after you fit thresholds. jev-latest and the marketing line jev-1.13 can move. Log the response model. TypeSafe’s published list price for jev-1.13 is $0.042 per million input tokens (vendor claim — confirm on the models page); output tokens are free on that same page. Unused distractors still bill as input.

Official Python and JavaScript SDKs read TYPESAFE_API_KEY and retry documented 429/529. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.

Pack map

FAQ

Does Jev execute the retrieval filter action? No. It returns typed answers. Your retrieval filter code calls queues, models, or humans.

Why several questions in one request? TypeSafe’s fan-out pattern: extra questions are cheap versus another HTTP call. TypeSafe’s parallel-questions cookbook: batch questions on one state. For many candidates, loop pairs (query, passage) or shortlist with BM25 first — official rerank cookbook — instead of one giant Choice over 200 ids unless you followed their line-search pattern.

Where is the rest of the RAG pack? Start with RAG input contracts and RAG confidence thresholds. Cluster hub: Use cases.

Should Jev generate the RAG answer? No. Classify or score passages; another model (or extractive code) writes. That is the cite-vs-generate boundary.

Do we publish rerank lifts? No. TypeSafe’s cookbooks may show measurements — treat those as vendor figures and re-run on your corpus.

What this page does not claim

• No invented top-1 / top-10 lifts.
• Not a vector database.
• Not official TypeSafe.
• Official TypeSafe status, or that jev.pro issues API keys.
• That a schema-constrained answer is automatically factually correct.

Disclaimer

This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai.

Primary documentation: https://docs.typesafe.ai. Hub: Use cases.