Skip to content
JevDirectory.org
Patterns

Guardrail

Screening messages going into or out of an LLM app with hazard Nouls and severity Scores, then thresholding the probabilities.

Open docs.typesafe.aiAll 117 terms
Category
Patterns
Also known as
—
Related terms
4
Directory entries
35
Docs
docs.typesafe.ai
Added
2026-09-24

Definition

The guardrails cookbook screens every message with one request: a Noul describes each possible hazard, a Score rates severity, and code thresholds the probabilities to pass, review, block, or route the message.

The RAG passage cookbook applies the same shape to retrieved context, using per-passage Nouls for relevance, evidence, contradiction, and injection before anything reaches the generator.

Tagspatternsguardrailssafety

Related terms

Definitions that connect to this one.

RAG passage classificationScore each retrieved passage with four Noul questions, then decide in code which passages reach the answering model.PatternsNoulThe yes/no primitive: one probability from 0 to 1, where the value is the answer and no separate confidence is returned.PrimitivesThresholdThe value your code compares against a score, a Noul probability, or a confidence to turn an answer into an action.Answers & ConfidenceAdversarial contentJev treats state as data, not as hostile input; injected instructions can move answers, so write precise criteria and test edge cases.Evaluation & Training

From the directory

Cookbook: Guardrails for LLMs
Screen every message going into and out of an LLM app: a Noul describes possible hazards, a Score rates severity, and code thresholds the probabilities.
Cookbooks & DemosDocs#official#cookbook#guardrails
Official
Cookbook: Classifying RAG Passages
Score retrieved passages with four Noul questions each, then route in code: prompt injections dropped, false premises kept as conflicts, and evidence assembled separately for the generator.
Cookbooks & DemosDocs#official#cookbook#search
Official
jevals
85GitHub stars
OpenLayer's eval and guardrail library that replaces LLM judges with Jev-style decisions, packing every check for an agent trace into one request.
Repos & SDKs#community#python#evaluation
Communityopenlayer-ai/jevalsPython
fx auto mode benchmarked with Jev
Vercel's Pranit reports benchmarking the fx auto-mode safety classifier with Jev: about 5 to 18 times faster and more accurate than GPT-5.6 Luna.
Practices & PatternsPost#community#x#benchmark
Community
TypeSafeAI for .NET
4GitHub stars
.NET SDK for the System One API with typed question sets, HttpClientFactory and dependency-injection wiring, plus Microsoft.Extensions.AI guardrail, routing, tool, and evaluator adapters.
Repos & SDKs#community#sdk#csharp
CommunityHawxy/TypeSafeAI.NetC#
TypeSafe Code Guard
2GitHub stars
An MCP server for Antigravity IDE, Cursor, and Claude Code that uses Jev (jev-latest) to locate code semantically and audit git diffs, replacing broad grep sweeps of entire files.
Tools & Integrations#community#python#mcp
Communitygreenyamao/Antigravity-mcp-semantic-search-with-TypeSafeAiPython

29 more matching entries in the full directory.

From the community

Posts from builders shipping with Jev right now.

Follow @typesafeai

An Elixir port of the official SDK

jod: Zod validates data, Jev judges it

Loki: a self-improving harness with a Jev companion

Jev as an LLM guardrail inside Agentgateway

Jev reranks station suggestions in TrainLCD

Vercel's fx safety reviewer, measured at 18× faster