Approxima

Approxima is an AI testing service that plugs into CI/CD and runs agents on every pull request. The agents build the app in an isolated sandbox, test it in a real browser, and report exactly what is broken, with screenshots. The company positions it as a layer on top of Playwright that debugs flaky end to end tests and heals them, so a red build signals broken code rather than a broken selector. It is aimed at engineering teams that rely on browser tests but do not want to spend time maintaining Playwright selectors or chasing flaky failures, and who are wary of moving their whole suite onto a proprietary third party platform. Work is organised as journeys and suites. A command line tool, installed from npm, lets developers log in, list journeys and run history, run a single journey or a whole suite, trigger a review of a given pull request, and stream progress live in the terminal. An explore command takes a goal in plain English, lets the agent navigate the app to discover the steps, and offers to save the result as a reusable journey. It requires a GitHub repository connected through the Approxima GitHub App. Approxima runs on credits, but no plan names or prices are published on the site; pricing is not public. The company's stated promise is that tests keep working exactly as before if credits run out or a team stops using the service, even when the Playwright wrapper is in place, which limits lock in.