Blog
Writing for teams that ship every day
Regression testing, bug reports that developers can act on, and what changes when most new code is written by a model.
2026-10-08How to write a test plan a model can executeA plan written for a person leaves out what the person already knows. A plan a model can run names the start, the steps, the data and the outcome, and nothing else.2026-10-08Visual regression testing without the noiseScreenshot comparison finds the breaks nobody wrote an assertion for, and drowns them in font rendering and animation. How to keep the signal.2026-10-08Testing a site behind a login or a preview gateMost real tests start with a sign-in and many run against a preview that is itself behind a gate. How to get the browser through without putting a password in a test.2026-10-08How to test what Claude Code (or Cursor, or Copilot) wrote before you merge itReading an agent's diff tells you what it meant; running the product tells you what it did. Here is a merge check that takes an hour to set up, runs on every change, and sits beside code review rather than replacing it.2026-10-08Snag vs Marker.io: a widget that files to your tracker, or a tracker built for the fixMarker.io is a feedback widget and extension with two-way sync to Jira, Linear and others. Snag is capture plus tracker, with mobile and an MCP for coding agents. The right one depends on whether your tracker is the centre of your world.2026-10-08Snag vs Jam: bug capture for developers, with or without the testsJam is the best-known developer bug recorder, with video, console and network, an MCP server and a CLI. Snag captures the same context with a screenshot or a short recording and a replayable trail, adds a triage board and mobile capture, and sits next to a test runner that proves the fix.2026-10-08Snag vs BugHerd: feedback pinned to a web page, or bugs with the evidence attachedBugHerd pins sticky notes on a live website for clients and agencies to review. Snag captures what a developer needs to fix a bug, on web and mobile, with a coding agent in mind. Different jobs, often confused.2026-10-08Self-healing tests: what works, what is marketingA test that quietly rewrites itself to pass is not healing, it is lying. What a repair should do, what it should never do, and the question to ask a vendor.2026-10-08Regression testing for AI-generated code: a practical guideWhen a model writes most of the diff, the review you used to rely on is gone. What takes its place, and how to set it up without a QA team.2026-10-08Playwright vs record-and-replay: when each winsHand-written Playwright and recorded tests are not rivals. One is for the code you own in depth, the other for the product everyone touches. Here is where the line falls.2026-10-08An Octomind alternative: where to take your generated Playwright testsOctomind shut down at the end of May 2026. Teams that liked auto-generated tests from a URL with a pull-request check can get the same shape, with a human-reviewed plan and proven tests, at a published price.2026-10-08Giving AI coding agents your bug tracker: MCP for QAAn agent that can read the bug, see the error and the failed request, fix it, and run the covering tests before it says done. What the connection looks like and what to keep out of its hands.2026-10-08FlowQA vs QA Wolf: a managed QA team or a tool your team runsQA Wolf sells coverage as a service, with people who maintain your tests. FlowQA sells the tool, at a published price, for teams that would rather own their tests. Which fits depends on budget and who you want holding the suite.2026-10-08FlowQA vs Momentic: two self-serve AI testing tools, comparedBoth publish prices, both ship a free tier and an MCP server, both generate and heal tests. They differ on where tests live, how healing is checked, and what sits next to the test runner.2026-10-08FlowQA vs mabl: enterprise low-code testing against a tool for small teamsmabl is a mature low-code platform sold to mid-market and enterprise QA, with prices behind sales. FlowQA publishes its prices and is built for teams with no QA function. Here is how they differ and who each is for.2026-10-08Flaky tests: causes and a triage methodA flaky test is one that fails, then passes, with no change in between. Five causes account for nearly all of them, and each has a different fix.2026-10-08Bug reports developers can act on: what to captureThe difference between "the checkout is broken" and a report that gets fixed in ten minutes is four pieces of context, all of which the browser already has.2026-10-08Autonomous test generation: how it works, where it fails, what to reviewAn agent can map a site, propose journeys, write the tests and prove them. It still needs a person at three points. Here is what happens at each and what to look at.