Cookbooks
Official TypeSafe cookbook recipes for guardrails, re-ranking, RAG passage classification, citation checks and more with Jev.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
TypeSafe's Cookbooks section of the Jev documentation collects worked recipes for common decision-model tasks. According to the list summary, recipes include LLM guardrails, re-ranking, classifying RAG passages, checking citations, and other patterns.
The cookbooks are a starting point for developers who want to see how to phrase state and typed questions for a specific job and how to act on the returned probabilities. They are documentation rather than a standalone project.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does this LLM output violate the guardrail policy?(inferred) | yes · no |
| Score | How relevant is this passage to the query?(inferred) | — |
| Yes / no | Does the cited source support this claim?(inferred) | yes · no |
How it works
- Pick a recipe matching your task
- Adapt the state and question definitions
- Call Jev and apply the recipe's threshold or ranking logic
Strengths
- Official, task-specific examples
- Covers guardrails, retrieval and citation checking
Caveats
- List-only entry; recipe contents not reviewed here
- Requires a TypeSafe key to run
Getting started
Start at https://docs.typesafe.ai/cookbooks/llm_guardrails.At a glance
- Type
- Article
- Status
- production
- Has code
- yes
Use cases and integrations
How curated lists describe it
Official recipes for guardrails, re-ranking, RAG passage classification, citation checks, and more.
MrJev/awesome-jev
Frequently asked questions
What recipes are included?
Guardrails, re-ranking, RAG passage classification, citation checks and more, per the list summary.
Who publishes them?
TypeSafe, as part of the official Jev docs.
Do I need an API key?
Running the recipes requires a TypeSafe key.
Related projects
MCP server giving coding agents typed judgment tools (verify, screen, rerank, gate, score) backed by Jev, with auto/review/escalate policy.
MCP serverbetaPythonSmall stdio MCP server with ask, rank and verify tools backed by Jev, a local event log and installers for several agent harnesses.
MCP serverbetaTypeScriptAgent skill that teaches coding agents to hand bulk semantic judgments to Jev and to find where a product could use Jev decisions.
Agent skillbetaWorked Jev examples for ticket triage and reranking, runnable through OpenRouter without an early-access key, with offline tests.
ProjectprototypePythonTypeSafe cookbook showing how to evaluate many typed questions over one shared state in a single Jev request.
ArticleproductionTyped-decision accuracy layer for RAG: six checks for intent, relevance, clarification, sufficiency, grounding and caveats.
ServicebetaPython