QualityPilot
Free tool · no sign-up

Jest flaky test detector

Paste a Jest or Vitest test file. Get a 0–100 risk score and per-pattern fixes in 2 seconds. Free, no signup.

1

Paste

Drop your test source into the box.

2

Scan

We check for 10 known flaky patterns (timers, races, network, etc.).

3

Fix

See your risk score plus actionable suggestions per pattern.

Max 50 KB. Your code is processed in memory and never stored. 10 free analyses per IP per hour.

Paste your test file above and click Analyze to see your risk score.

Why flaky tests matter

A flaky test is one that passes and fails for the same code, depending on load, ordering, or timing. Each re-run costs CI minutes (real money), erodes trust in your suite, and trains engineers to ignore red builds. At even a small team scale, a single chronically flaky test can burn ten engineer-hours a month in re-runs and triage. Catching the pattern early — before it ships — is cheaper than debugging it in production.

About QualityPilot

This tool is a one-file slice of the QualityPilot scan. The full product runs the same heuristic across every test in your repo, scores it A–F, cross-references your CI history, and proposes AI fix-PRs for the worst files.

See it on every PR →

FAQ

Do you store the code I paste?

No. The analysis runs in-memory, the source is never written to disk or any database. Logs only contain the risk score, not your code.

What patterns do you detect?

We scan for 10 categories of known flaky-test smells:
  • Fixed sleep/timeout
  • Network / external API
  • Database / ORM
  • File system I/O
  • Email / messaging
  • Date/time dependency
  • Random / dynamic data
  • Concurrency / race conditions
  • UI animation / transitions
  • Shared state / order dependency

Each detection requires both a keyword signal and a structural signal (an actual import or function call) — that's why a test that just mentions “timeout” in a comment doesn't false- positive.

How is this different from the full QualityPilot scan?

This tool is heuristic-only and instant. The full scan additionally runs an AI fix proposal on the worst test, scores against your repo's actual CI run history, and opens fix PRs.