CueTest

CueTest is a coding tool. CueTest runs end-to-end browser tests written in plain English on hosted browsers, verifies each expected result independently, and returns traces, screenshots and reports. It plugs into CI pipelines and starts free.

CueTest is a hosted platform for end-to-end browser testing in which tests are written as ordinary sentences instead of selector-based code. Each test is a sequence of ordered steps, and every step carries one instruction and one expected result. Teams can reuse step groups, run a selected range, or execute a whole project in order, with credentials referenced through environment placeholders rather than typed into steps. Execution happens in real hosted browsers, with one browser context kept for the full run. An AI agent performs each step inside a fresh, bounded context, and a separate verifier then checks the expected result against the live page. That split matters: a step passes only when the page confirms the outcome, not when the model claims it finished. Successful interactions are saved as a replay plan. Later runs try that deterministic plan first and call the AI only when the saved path no longer proves the result, which is the healing behavior that keeps maintenance low when an interface changes. Every run returns evidence: a trace, screenshots, a report, and the exact failed step. Runs can be triggered from GitHub Actions, GitLab CI, or any pipeline able to call an API, so the output can gate a release. The homepage also offers a no-signup hosted trial run against a public URL, limited to one run per browser window and three minutes. The audience is fast-moving SaaS teams, lean engineering-led QA groups, and teams already running Playwright or Cypress who want adaptive coverage for the journeys that change most, such as checkout, onboarding and permissions. It is positioned as a complement to coded suites rather than a full replacement. Pricing is based on browser minutes and AI resolutions instead of the number of saved steps, which sets it apart from tools that charge per test. Alternatives include maintaining Playwright or Cypress scripts by hand or using other AI testing agents, and CueTest's distinguishing choices are independent verification and replay-first execution.