Deployment regression comparator
Compare locally supplied captures from the same owned-build route, HTTP and visible-UI fixtures, retaining failures and unknown observations.
Compare two owned-build captures
Run the same fixtures against your baseline and candidate builds, then paste their capture files below. Comparison stays in this page. It makes no test requests and uploads neither capture.
Get the capture runner, fixture template and setup instructions. The runner performs real HTTP and browser checks on a target you explicitly acknowledge owning. Use self-contained, unauthenticated builds; blocked external requests make UI coverage incomplete.
Each visible-element fixture requires exactly one matching element. Optional text is compared after whitespace normalization. A changed response digest is a difference to inspect; a newly failed expectation is reported separately.
Up to 8 routes and 8 UI checks per route; 64 KiB per capture. Inputs and results clear on reload and page lifecycle transitions.
Keep the comparison meaningful
Use identical fixture files, viewport and Chromium version for both captures. Pages with time-dependent or personalized content may legitimately change their response digests. The runner observes HTTP responses and UI separately; it does not establish an atomic release snapshot or run clicks, logins or destructive actions.
Exports contain anonymous fixture numbers, statuses, counts and comparison outcomes, without raw URLs, selectors, response bodies, captured text, cookies or credentials. Imported files can be edited by their supplier and are not independently authenticated.