// regression testing for SaaS
Regression testing that keeps up with your release cadence.
Stop spending a day a week clicking through your own app. AegisRunner crawls every page, writes Playwright tests, runs them on every deploy, and auto-heals the suite when your UI shifts.
// the problem
What "regression testing" usually looks like in a fast-moving SaaS team.
Manual regression takes a full day
Every release means 4–8 hours of click-through QA. Bugs slip when the team is tired or rushed.
Hand-written E2E tests rot
Class names change, layouts shift, selectors break. The suite goes red on green builds; everyone stops trusting it.
No coverage of the long tail
You test the checkout, you test the sign-up. The 200 other pages that subtly broke last Thursday? Nobody caught it.
QA hires don't scale
One person can keep ~30–50 hand-authored Playwright tests green. Past that, maintenance overruns new test creation.
// our take
A regression suite that writes and maintains itself.
Crawls every page from a URL
No sitemap needed. We walk every link, every form, every modal — including auth-gated pages once you give us a login.
Generates Playwright tests in minutes
AI grounds tests in your actual DOM, picks resilient role/text/data-testid locators, exports clean TypeScript.
Auto-heals on UI drift
When a button gets renamed or moved, AI repairs the selector and re-runs the test before paging anyone.
Runs on every deploy
Webhooks from Vercel / Netlify / GitLab fire the suite the moment your release lands. CI integration on every paid plan.
Pages you only when something is broken
Slack / Discord / Jira / GitHub Issues with screenshots, locators, and Playwright trace attached.
Bundles a11y / SEO / security audits
Every scan also runs axe-core (WCAG 2.1 AA), Lighthouse SEO, and security-header checks. One scan, four signals.
// faq
Frequently asked
How is this different from writing Playwright tests by hand?
It's the same engine — AegisRunner runs on Playwright under the hood. The difference is who writes and maintains the tests. With Playwright direct, you assign engineering hours; with AegisRunner, the AI generates, runs, and auto-heals the suite. You can also export the underlying Playwright code on every plan and take it with you.
What about flaky tests?
AI auto-heal repairs locators that drift, and the verification pipeline re-runs every generated test against the live site before saving — so brittle selectors get filtered out before they hit production. Our internal benchmark is < 1% flakiness on a 50-test SPA suite over 10 consecutive runs.
Can we use this on staging behind basic auth?
Yes. Configure a login script (or upload pre-auth cookies) under Project Config → Test Environment. We support form login, OAuth flows, and bypass tokens for WAF-gated staging environments.
How fast does a scan finish for a 100-page SaaS?
Typically 4–8 minutes for the crawl + 1–2 minutes for AI generation. Subsequent re-scans on stable pages use replay-cache and finish in roughly half the time.
Try it on your SaaS — no card
First scan in under a minute. You get the report, the Playwright suite, and the audit results on the same page.