I built TrailWrightQA to let developers, QA teams, or business-analysts generate browser UI tests without writing code. It runs locally and requires an API key from OpenAI / Gemini / or Anthropic. All test code and data remain on your machine — no external servers beyond the LLM call.

Because it’s open-source and self-hosted, each test run is free (beyond the LLM cost). That eliminates recurring per-run fees typical of many automated-testing services.

It’s rough around the edges, especially for complex UI flows — but I'm limited on dev time and so releasing it publicly in case others have similar needs or want to contribute. Feedback, pull-requests, or issue reports welcome.

Repo: https://github.com/marktl/TrailWrightQA

— Mark