SAVED DEMO PROOF · REAL FILES
Ten release blockers found. One bounded patch cleared every recorded blocker.
AtlasPay is a bundled multilingual fixture, not a customer. The run below happened locally and saved its reports and screenshots. This page shows those files exactly as recorded; opening it does not rerun or invent a result.
- 10release blockers
- 3approved files changed
- samechecks run again
- 0release blockers left
- PASSen-GB source locale
The run record
| Baseline scan | atlaspay-83210b0daea2 |
|---|---|
| Verification scan | atlaspay-75123a28884f |
| Mode | genuine deterministic replay artifact · no-AI deterministic |
| Routes | / · /pricing · /dashboard · /checkout · /settings |
| Locales | hi-IN · de-DE · ar-SA · he-IL · ja-JP · zh-Hans-CN · th-TH · fr-FR · es-MX · en-GB |
| Viewports · browsers · themes | mobile 390×844, tablet 768×1024, desktop 1440×900 · chromium · light, dark |
| Recorded | 2026-08-03T13:36:31.484Z → 2026-08-03T13:36:31.621Z |
| Proof generated | 2026-08-03T13:36:36.582Z |
What failed, by rule
- 1vertical clipping
- 1cta overflow
- 1wrong direction
- 1rtl icon order
- 1raw translation key
- 1font coverage
- 1line breaking
- 1placeholder mismatch
- 1glossary violation
- 1wrong page lang
10 recorded issues across 5 routes and 10 locales. Each one stores the measurement taken and the predicate it failed; the workspace shows them against the fixture itself.
The patch
apps/demo-target/data/layout.jsonallowlisted for repairapps/demo-target/data/locale-state.jsonallowlisted for repairapps/demo-target/data/translations.jsonallowlisted for repair
The repair could touch nothing outside that list. The diff itself is published at /replay/repair.patch.
Verification
10 → 0verified with warnings
- Verified at
- 2026-08-03T13:36:36.582Z
- Source-locale regression
- PASS
- New blocking issues
- 0
- Console error delta
- not measured
- Accessibility regression
- not measured
- Diff within policy
- true
Recorded as not measured on this path, so not readable as a pass: consoleErrorDelta, accessibilityRegression.
Portable exports the run wrote
- /replay/report.jsonJSON report
- /replay/report.htmlHTML report
- /replay/report.sarifSARIF
- /replay/junit.xmlJUnit XML
- /replay/issues.csvCSV issues
- /replay/screenshots.zipScreenshots ZIP
- /replay/repair.patchUnified patch
- /replay/repair-proof.jsonProof JSON
The post-repair exports are deliberately empty: after the patch there was no issue left to list.
What this proves and what it does not
- It proves the engine detects, repairs and reverifies a defined set of localisation defects with the identical predicates.
- It is a fixture, so it says nothing about your site. For third-party targets, the published external scans are the honest comparison.
- Fields the run did not measure are published as not measured. They are never rolled into the verdict.