Studio guide / Evidence

Turn a failed run into a test that stays failed until it is fixed.

Studio keeps comparison, failure evidence, replay, and regression promotion connected to the same versions and source runs.

Evidence path

Preserve the chain instead of copying a screenshot.

  1. 01
    Locate

    Select a baseline and candidate and identify the first sustained meaningful divergence.

  2. 02
    Inspect

    Review the failure, supporting events, artifacts, scope, and missing evidence.

  3. 03
    Replay

    Reproduce the condition with the pinned environment, scenario, system versions, and seed.

  4. 04
    Promote

    Create a regression case only when the reproduction contract is complete.

NO AUTOMATIC CAUSAL CLAIM

A comparison locates behavioral change. It does not claim that the first differing signal caused the failure unless a controlled design and its evidence support that conclusion.