Home /Benchmarks, evals and calibration /jevals
jevals
Runs a trace's agent, quality, and security evals as one batch of typed Jev questions instead of separate LLM-judge calls.
From the readme
jevals Evals and guardrails for agents, using Jev-style decision models instead of an LLM judge. All the evals for a trace go out as one request that costs a few thousandths of a cent and comes back in a few hundred milliseconds, so you can run them on every trace and inside the agent loop. Works with Jev through the TypeSafe or Vercel APIs, with any …
Details
- Section
- Benchmarks, evals and calibration
- owner
- openlayer-ai
- stars
- 111
- forks
- 12
- pushed
- 2026-11-02
- Language
- Python
- License
- MIT
What it is
- Form
- Library or SDK
- Audience
- Agent builders
Maturity
Docs
reports measured numbers