The Radar
Drop in a few runs, get the picture
Drop JUnit XML files here
One file per CI run — add 3+ runs for meaningful flakiness signal. Also supports surefire/pytest/Playwright JUnit output.
Flaky tests — ranked by score
Done ✓
| # | Test | Flakiness | Runs (pass/fail) | Likely cause |
|---|
A test failing every run is broken, not flaky — FlakeRadar separates them so you fix the right thing. Always-pass tests aren't shown.
Method
How the score works
1
Flip rate
How often a test switches pass↔fail between consecutive runs — the core instability signal. Pure alternation scores highest.
2
Balance & failure ratio
A 50/50 pass-fail split is flakier than a 1-in-10 blip; a 10/10 fail is broken, not flaky (scored separately).
3
Root-cause taxonomy
Error text is matched against an 8-category taxonomy — timing, order-dependency, data collision, environment, and more.
4
Quarantine & fix
Export a ready-to-paste quarantine list to green your pipeline today, then fix by category with the PRO RCA guidance.
Pricing
Free to scan. $29 to fix like a pro.
Free
$0
- Unlimited in-browser analysis
- Flakiness scoring & ranking
- Broken-vs-flaky separation
- Basic cause hints + CSV export
Founder License
$29 one-time · ~₹1,499
- Full 8-category RCA engine (cause · why · fix)
- Quarantine-list export (Playwright/JUnit/pytest/Cypress)
- The flaky-test taxonomy reference
- JSON export + trend across sessions
- Lifetime license, this site
CI-Integrated
$79 /month
- Auto-ingest from GitHub Actions / Jenkins
- Hosted trend dashboard + history
- Weekly Slack / email digest
- Post-launch roadmap — join the list