What it does
LitmusRun runs AI agents against a web app link in Chromium. The link can be a staging or production URL. You write a mission in plain language, such as completing checkout as a first-time buyer. You can also add audience notes. The agents follow the path you define and record what they click. The target can be a full product, one feature or a custom path. Each run produces session evidence, a bug log, heatmaps and panel consensus. Every item links to a timestamp. The research report gives a ship or don't-ship verdict with ranked findings. Each finding links to step screenshots.
Capabilities
- Session evidence logs clicks, rage events and hesitation.
- Export packs come as Markdown, JSON or an agent-ready fix list.
- Encrypted unlocks store keys encrypted at rest. LitmusRun says the browser never sees the values again.
- Private screenshots are delivered as authenticated media with signed URLs.
- Live progress shows exploration milestones as they happen.
When to use it
- Check a staging link before launch.
- Test a defined user path, such as signup to pricing to pay, or completing checkout as a first-time buyer.
- Rerun the same link and mission to track regressions, using cross-run memory on Business+ plans.
