Mendral

Mendral is an AI-powered DevOps engineer that continuously monitors CI/CD pipelines, diagnoses build failures, detects flaky tests, and automatically generates fixes. Built by the team behind Docker and Dagger, it installs as a GitHub App and starts providing insights within hours, requiring no infrastructure management or configuration files.

• CI Failure Diagnosis - When a build fails, Mendral reads the full log output, identifies the root cause, and provides an actual explanation with a confidence score rather than just a stack trace dump.

• Flaky Test Detection - Correlates failures across hundreds of CI runs to find real causes like timing dependencies, shared state, and order-dependent execution, then traces flakes back to the commit that introduced them.

• Automated Fix Generation - When confidence is high enough, Mendral opens a pull request with the fix that you can review like any other PR from your team.

• Smart Notifications - Mentions the people who actually have context: the author of the broken commit, whoever last touched that code, or known maintainers. Known flakes don't trigger pings.

• Noise Reduction - Groups related failures, waits for patterns to emerge, and surfaces what actually matters, such as identifying which tests caused the majority of re-runs in a week.

• Pattern Recognition - Builds hypotheses and refines them as more data comes in. A test failing once is a data point; the same test failing across multiple unrelated PRs becomes a pattern that Mendral connects automatically.

To get started, install Mendral as a GitHub App. The setup takes about five minutes with no infrastructure to manage. It immediately begins reading your CI logs and typically delivers the first actionable insight within hours. Mendral integrates with Slack for team notifications and is currently managing CI/CD in production for teams including PostHog, Inngest, Anyshift, Blaxel, Oway, and Golf.