Web Performance Regression Tester

Performance regressions are one of the most insidious quality problems in web development — they happen gradually, they are invisible until users complain, and by the time they are noticed, the responsible change is buried under dozens of subsequent commits. Setting up automated performance regression testing means catching these regressions at the pull request stage, before they ever reach production. This AI assistant helps development teams build exactly this kind of performance safety net.

The assistant specializes in configuring and interpreting automated performance testing pipelines. It sets up Lighthouse CI to run audits on every pull request, comparing Core Web Vitals — Largest Contentful Paint, Cumulative Layout Shift, Interaction to Next Paint, Time to First Byte — against defined budgets and failing builds that exceed thresholds. It generates the `lighthouserc` configuration, CI workflow files for GitHub Actions or GitLab CI, and the assertion rules that define what constitutes an acceptable performance regression.

Beyond Lighthouse CI, the assistant covers WebPageTest API integration for more realistic performance measurement under controlled network and device conditions, `bundlesize` and `size-limit` for JavaScript bundle budget enforcement, and Webpack Bundle Analyzer or Rollup plugin visualizer for understanding what is contributing to bundle bloat. It helps you define per-route performance budgets that reflect your users' device and network distribution.

The assistant also covers performance debugging workflows: identifying what caused a regression by bisecting builds, attributing bundle size increases to specific dependencies or code changes, and correlating Lighthouse metric changes to specific rendering or loading behaviors. It explains each Core Web Vital in terms of what user experience it captures and what technical factors most commonly cause it to regress.

This assistant is ideal for engineering leads establishing performance standards for their team, frontend developers investigating a performance alert, and DevOps engineers adding performance gates to CI/CD pipelines. Expect CI configuration code, budget definition files, analysis of Lighthouse reports, and actionable remediation suggestions.