// regression testing for SaaS
Regression testing that keeps up with your release cadence.
AegisRunner explores reachable product paths, writes and reviews test cases, runs them on your release cadence, and keeps failures tied to the page and evidence that produced them.
// 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
As coverage grows, test data, selectors and environments can compete with new test creation for engineering time.
// our take
AI-native regression testing with reviewable maintenance.
Explores reachable product paths
A sitemap is optional. The crawler follows links and interactive states within the scope and page budget you set, including authenticated paths when login succeeds.
Generates grounded Playwright tests
AI uses the observed DOM and prefers role, text and data-testid locators; workspace members can export supported web tests as TypeScript.
Auto-heals on UI drift
AI can repair eligible locator drift and re-run the case. Changed assertions and inconclusive outcomes remain visible for review.
Runs on every deploy
Webhooks from Vercel / Netlify / GitLab fire the suite the moment your release lands. CI trigger automation is available on Pro and Business plans.
Failure evidence for your team
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.
Covers your mobile app too
Ship an iOS or Android app alongside the web app? Upload the build — AI explores it on a real device, generates the tests, and runs them on the same schedule.
Cross-platform journeys
One test creates a record in the mobile app and verifies it on the website — regression coverage for cross-surface sync paths.
// related guides
Plan a regression strategy that lasts.
// faq
Frequently asked
How is this different from writing Playwright tests by hand?
It's the same browser engine, with a different operating model. AegisRunner discovers paths, generates and reviews cases, runs them, and can repair eligible locator drift. Standard Playwright export is available to workspace members; POM project export requires Pro or Business.
What about flaky tests?
AI triage can propose repairs for eligible locator drift, while runtime evidence shows which cases were actually exercised. Application state, test data, network behavior and business assertions can still produce flaky or inconclusive results, so the run is never presented as silently green.
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?
It varies with authentication, reachable states, page budget, model latency and the depth of generated testing. Use a representative scan to size CI timeouts rather than relying on a universal estimate.
Can it regression-test our mobile app as well?
Yes. Upload your .apk or .ipa and AegisRunner explores the app on a real iOS or Android device, generates test cases the same way it does for web, and runs them on real hardware. Cross-platform journeys go further: a single test can act in the mobile app and verify the result on the website, catching sync regressions between the two. Mobile testing is included on Pro and above.
Try it on your SaaS — no card
Start with a representative page budget. You get the report and generated cases together; Pro and Business plans add Playwright/POM export.