# Jevals – replacing LLM judges with typed Jev decisions

> Agent evals and guardrails in one request. Built on Jev, Kev and Laya. - openlayer-ai/jevals

Page: https://nowfound.app/launch/jevals-replacing-llm-judges-with-typed-jev-decisions

- Launched: 2026-09-20
- Launch-day votes: 47, 6 comments; #691 of 8,095 launches that month (top 9%)
- Website: [github.com](https://github.com/openlayer-ai/jevals), online when checked 2026-09-20
- Category: [AI](https://nowfound.app/c/ai.md)

## In plain words

Jevals is a tool for evaluating and protecting AI agents by replacing traditional language model judges with typed decisions. It combines agent evaluation and guardrail functionality in a single request, built on the Jev, Kev, and Laya frameworks. The software is designed for developers working with AI agents who need to assess performance and implement safety constraints simultaneously through a structured decision-making approach.

## Does a similar job

- [Jev](https://nowfound.app/launch/jev): Fast, structured AI decisions for software automation (2026-09-21, 544 votes, console.typesafe.ai)
- [Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks](https://nowfound.app/launch/forge-guardrails-take-an-8b-model-from-53-to-99-on-agentic-t): Hi HN, I'm Antoine Zambelli, AI Director at Texas Instruments. I built Forge, an open-source reliability layer for self-hosted LLM tool-calling. What it does: - Adds domain-and-tool-agnostic guardrai… (2026-05-19, 687 votes, github.com)
- [Agents.json – OpenAPI Specification for LLMs](https://nowfound.app/launch/agents-json-openapi-specification-for-llms): Hey HN, we’re building an open specification that lets agents discover and invoke APIs with natural language, built on the OpenAPI standard. agents.json clearly defines the contract between LLMs and… (2025-03-03, 212 votes, github.com)
- [Venn.ai](https://nowfound.app/launch/venn-ai): Delegate real work to AI agents with safety guardrails (2026-03-26, 336 votes, venn.ai)
- [Jevstiller – Distill Jev into a local model, with a disagreement bound](https://nowfound.app/launch/jevstiller-distill-jev-into-a-local-model-with-a-disagreemen): September 2026. Every number here is from the benchmarks, and bash experiments/bench.sh --no-record reruns them without an API key. (2026-09-29, 67 votes, jevstiller.pages.dev)
- [APIEval-20](https://nowfound.app/launch/apieval-20): An open benchmark for AI agents that test APIs (2026-05-08, 121 votes, resources.kusho.ai)

All alternatives: https://nowfound.app/alternatives/jevals-replacing-llm-judges-with-typed-jev-decisions.md
