Blog
Announcements and deep dives from the TestRelic team.
Ask AI, Your Choice of Model: Auto, DeepSeek and Sonnet 5
Ask AI now has a model picker (Auto, DeepSeek V4 Pro, DeepSeek Flash, Sonnet 4.6 and Sonnet 5), an organization-wide engine policy for admins, answers that name the model that wrote them, and a Remote tab for following Studio and CLI conversations live.
Which Test Is Which? Renames, Rewrites, and Tests That Share a Name
TestRelic now tells a test's name, its steps and its identity apart: a renamed test is recognised, a rewritten one gets a new version, tests sharing a name are called out, and Studio keeps same-named twins in separate rows.
Visual Regression Testing: From a Pixel Diff to the Element That Caused It
TestRelic now reports Playwright's visual comparisons, names the DOM elements behind a pixel diff, and lets you accept or reject a new baseline from the CLI, Studio or the cloud, with one record of who decided what.
Brittle selectors, dynamic elements, and the real cost of test maintenance
Test maintenance is where automation ROI goes to die: brittle locators, generated class names, and SPAs that re-render everything. A guide to selector strategies that survive UI change — and healing workflows for when they don't.
How to fix flaky Playwright tests: find the real cause, stop the retry roulette
Flaky tests are the #1 automation testing problem. A practical guide to the four root causes of flakiness — timing, isolation, environment, and data — and how cross-run history separates flaky tests from genuinely broken ones.
Slow test suites, slow feedback: making CI test runs worth waiting for
When the E2E suite takes 40 minutes, developers stop waiting for it — and quality signals arrive after decisions are made. How to find what's actually slow with duration trends, parallelize safely, and run the right tests instead of all of them.