By confident-ai

deepeval

Turn AI failures into rerunnable evals and targeted fixes.

LLM evaluation

Only the most recent Revisions are kept. Restoring one saves its content again as the newest.