TesterArmy e2e
8.1kA test step is a sentence like "upgrade the workspace to the Pro plan", and an agent finds the buttons itself on a website, an iPhone or an Android app. Successful actions are recorded and replayed later without calling the model. Open source, Apache-2.0.
We have not run it ourselves: described from the docs and the repository.
In numbers
What it is
A classic end-to-end test knows exactly what it verified, and you pay for that precision by rewriting a selector every time someone renames a button or changes a user flow. e2e from TesterArmy lets you describe the goal of a step in words, agent.act('upgrade the workspace to the Pro plan'), and the agent works out the path on its own, so a reworked page is far less likely to break the test. The same test keeps ordinary locators and expect calls, the way Playwright does, so the things that matter are still checked precisely.
What sets it apart from Playwright MCP, its neighbour in this catalogue, is the job. There an agent gets a browser to do work; here it gets a test suite you run before every release. Hence the cache: a confirmed action is replayed next time without the model, so your token spend follows how often the UI changes rather than how often the tests run. It suits anyone building a site or an app with an agent who wants checkout and sign-up clicked through automatically before every release.
How it works
npx e2e run.act carries out a goal described in words, assert checks a condition in words, extract pulls data off the screen, waitFor waits for a state. Plain expect(screen.getByRole(...)) sits alongside.assert, extract, waitFor) call the model every time.npx skills add tester-army/e2e installs a skill and an MCP server for Claude Code, Cursor and Codex. And e2e explore clicks through the app hunting for bugs before a PR.
Comments