Features
- JSON dataset of prompt test cases
- deterministic mock model so CI is stable
- scoring for structure, relevance, safety and actionability
- failure threshold and CLI output for pipelines
- browser dashboard
AI Eval Lab: prompt regression testing with CI gates
Calling an AI API is easy. Knowing that this week's prompt is not worse than last week's is not. AI Eval Lab treats prompts like code: they get tests, scores and a red build.
4quality dimensions
1deterministic mock model
CIgated releases
1browser dashboard