flaky-test-detection

Open-source Playwright reporter + CLI (on npm) comparing test runs to tell broken tests from flaky ones. Cross-run classification · root-cause rules with suggested checks · failures grouped by shared cause, so you fix once not nine times · in-report triage · spec-file view · trends · CI gate · HTML/JSON/Markdown. Fully offline, no AI.

GitHub Action that tells you whether a failing test is your fault, a known flaky test or already broken on main. No server, no dependencies.

TestSentry uses AI to analyze failed tests and provide failure categories, explanations, confidence scores, and suggested fixes. Each failure receives a unique fingerprint. The first request calls the AI model; repeated requests use the cached result, reducing duplicate calls, improving speed, and lowering AI usage costs.

Find flaky tests in any language - pytest, Jest, go test, JUnit. Local CLI, no SaaS, no account.

Root-cause diagnosis for flaky tests — identify timing sensitivity, order dependency, parallelism issues, and more.

Flaky test detection for GitHub Actions — tells flaky tests from real regressions using JUnit XML, test history, failure fingerprinting and diff blame, with optional LLM triage.

Self-hosted flaky-test detection for CI: JUnit XML ingestion, statistical flakiness scoring, GitHub issue automation