FlakeTriage

FlakeTriage is a web app that plugs into your CI (GitHub Actions, GitLab CI, Jenkins) and focuses on one ugly regression-testing reality: flaky tests and noisy failures. It ingests test results, retries, timing, and environment metadata, then clusters failures by signature and calculates a “flake likelihood” score per test and per suite. The app surfaces the smallest set of tests causing the most reruns, developer interruptions, and release delays. It also tracks regression risk by highlighting tests that recently became unstable after specific commits, dependency bumps, or infrastructure changes. This is an AI + traditional app: AI helps classify failure patterns and suggest likely root causes (timing, race conditions, data coupling, external services), but the core value is reliable analytics and workflow integration. It’s realistic: it won’t magically fix tests, but it will stop teams from drowning in red builds.