FlakyWatch
flakywatch.dev
What it is
FlakyWatch monitors GitHub Actions test results, classifies unstable tests, estimates CI and developer time costs, and tracks flakes as GitHub issues. Users upload JUnit XML through the FlakyWatch GitHub App and a workflow step; the service says agents and configuration files are unnecessary. Supported runners include pytest, Jest, JUnit, RSpec, and other JUnit-compatible tools. Tests receive stable, flaky, broken, or newly unstable labels with confidence scores. Generated GitHub issues include classification, impact, and recommended next steps, and can reopen when a flake returns. Dashboard metrics include blocked runs, main failures, affected branches, and cost per test. An optional weekly email digest reports new unstable tests, cost trends, and open issues. The free plan costs nothing and covers one repository with 14-day history. Starter costs $29 a month for five repositories and 30-day history; Team costs $79 a month for 20 repositories and 90-day history. Slack and CSV export are marked coming soon on Team.
Spec sheet
- What it does
- FlakyWatch monitors GitHub Actions test results, detects and classifies flaky tests, estimates CI and developer time costs, and tracks them as GitHub issues.Read 3 Oct 2026 · flakywatch.dev
- Integrations
- The site lists GitHub Actions and GitHub issues; Slack integration is marked coming soon on the Team plan.Read 3 Oct 2026 · flakywatch.dev
- Security
- The privacy policy states that data is encrypted in transit and at rest, production systems have access controls, and the service uses least-privilege access and regular security audits.Read 3 Oct 2026 · flakywatch.dev
- Support
- FlakyWatch lists email support on Starter, priority support on Team, and says it aims to respond to inquiries within 48 hours during business days.Read 3 Oct 2026 · flakywatch.dev
- Classification
- Tests are classified as stable, flaky, broken, or newly unstable, with confidence scores.Read 3 Oct 2026 · flakywatch.dev
- Data access
- The privacy policy says FlakyWatch accesses repository metadata, GitHub Actions workflow run data, uploaded JUnit XML artifacts, commit SHAs, branch names, and run timestamps, but not source code or other repository files.Read 3 Oct 2026 · flakywatch.dev
- Data retention
- Test execution data is retained for 30 days by default; aggregated statistics are retained for 90 days on paid plans and 14 days on the free plan.Read 3 Oct 2026 · flakywatch.dev
- Issue tracking
- Flaky tests can become GitHub issues with classification, impact, and recommended next steps; issues can be reopened if the flake returns.Read 3 Oct 2026 · flakywatch.dev
- Reporting
- The dashboard includes impact metrics such as blocked runs, main failures, affected branches, and cost per test.Read 3 Oct 2026 · flakywatch.dev
- Service limits
- The terms state that the free plan is limited to one repository and 14-day data history, and that monthly subscriptions are non-refundable.Read 3 Oct 2026 · flakywatch.dev
- Setup
- The site says setup uses the FlakyWatch GitHub App and a workflow step to upload JUnit XML artifacts, with no agents or config files required.Read 3 Oct 2026 · flakywatch.dev
- Test results
- Users upload JUnit XML results from GitHub Actions; the site names pytest, Jest, JUnit, RSpec, and any JUnit-compatible runner as supported.Read 3 Oct 2026 · flakywatch.dev
- Weekly digest
- FlakyWatch offers an opt-in weekly email digest with new unstable tests, cost trends, and open issues.Read 3 Oct 2026 · flakywatch.dev
Plans and prices
Strong points and limits
Strong points
- Classifies tests with confidence scores
- Creates and reopens GitHub issues
- Tracks CI and developer time costs
- No agents or config files required
- Free plan includes weekly digest
Limits
- Free plan limited to one repository
- Free history lasts 14 days
- Slack integration is coming soon
- Monthly subscriptions are non-refundable
Who it suits
FlakyWatch suits development teams using GitHub Actions that want classification, cost visibility, and issue workflows for unstable tests. It supports several common test runners through JUnit XML uploads.
Verdict
FlakyWatch combines flaky-test detection with cost reporting and GitHub issue lifecycle management. Plan limits mainly differ by repository count, history duration, and support level.
Where Specifiction places it
- Test Flakiness Management Software#15 of 317.8