QA Wolf

QA Wolf runs end-to-end testing as a service built on agents. Agents explore your application and map the test cases it needs, write each test in open-source Playwright (web) or Appium (mobile), and keep them in a suite you can read and export. On every pull request they pick the tests the change touches and run them in parallel; every failure is investigated by an agent that reproduces it and decides whether the app broke or the test went stale, and stale tests are repaired automatically as the UI changes. Real bugs arrive as reports. A recent addition (October 2026) lets you record a flow by clicking through it in a live browser, which the agent turns into a test with the same page-object code as the rest of the suite.

Coverage spans Chrome, Firefox and WebKit, Electron apps, iOS on real iPhones and iPads, Android on emulators or real devices, and native Windows and macOS apps. An MCP server and live-streamed runs connect it to coding agents and CI.

Pricing is per test run rather than per seat or per AI credit: per web run and per mobile or desktop run, for up to five minutes of run time, with each run retried up to three times at no charge to rule out flakes; volume pricing is available. It suits teams that want broad regression coverage without staffing a QA automation team.