Register the package once. Future uploads inherit build history, baseline, and saved test credentials.
- Package and version history stay together
- Verification feedback carries forward
- One personal workspace, with multi-workspace support on Pro
How it worksReview the workflow.
Review the workflow.
See what still must be proven.
Caraxe captures evidence first, ranks issues second, and keeps the final release call visible and reviewable.
deep review Export PDF
Do not release this build.
Two critical workflow gaps need attention before this release is safe.
F-01
Primary release gap
Validation accepted empty submit
criticalsignup workflow88% confidence
Open signup, submit empty fields, observe success state instead of validation feedback.
Issues122 critical
Evidence 76linked artifacts
Coveragepartialconfidence medium
The five-step flow
From APK upload to release decision.
Drag and drop a build, or trigger a review through API and CI when a release candidate is ready.
- Manifest validation before review
- Version metadata extracted automatically
- Poll the status endpoint for the verdict
The runner covers onboarding, auth, core workflow, settings, and a secondary flow when budget allows.
- Deep release reviews explore broadly and refresh the baseline
- Replay reviews compare quickly against baseline
- Workflow coverage is tracked explicitly
Caraxe identifies what changed since the golden build and separates new issues from existing debt.
- New, recurring, resurfaced, escalated, and resolved states
- Stable finding identity across runs
- Aggregate counts stay policy-neutral
The report opens with the evidence-supported assessment, top fixes, coverage, and available reproduction context.
- Release, hold, block, or inconclusive outcome
- Top issues ranked by release risk
- Calibration feedback sharpens future reviews
What the report answers
Four questions. One report.
Did the required flows pass?
Auth, signup, core workflow, and settings are scored with evidence.
What are the top fixes?
Priority uses severity, impact area, recurrence, and confidence.
What changed vs. baseline?
New, escalated, resurfaced, and resolved findings are tracked per issue.
What evidence supports this?
Report findings link to screenshots, the action trace, and - once a baseline exists - the build-over-build diff.
Help validate it on real release work.
Private validation is in progress. We are accepting a small number of Android teams whose apps fit the current supported scope; public access waits for multi-app certification.