Introduction

Scout is self-healing browser QA. You describe a test in plain English; Scout verifies it in a real browser, records a deterministic path, and replays that path for free on every later run.

The bottleneck moved

Code used to be the constraint. Agents now write it on demand, so code is the cheap part. What stays scarce is judgment: does the product actually do what you pictured?

Scout is the quality gate for that question. You state what “done” looks like in one sentence, and Scout proves the product clears that bar — then re-checks itself, unattended, when the interface drifts.

How a scenario runs

1. Describe. Write one plain-language scenario. There is no test file to own or maintain.
2. Verify. On the first run, an agent drives a real browser and confirms the outcome.
3. Record. Scout captures the path it took as a deterministic script.
4. Replay. Later runs follow the recorded path without an agent.
5. Heal. When the interface moves and a step breaks, Scout repairs the path at runtime — no human in the loop.

Where it fits

Scout is built for developers and AI coding agents adopting browser QA into autonomous loops. The scout report --check command exits non-zero on any non-verified scenario, so it slots directly into CI as the gate for an automated fix-to-PR cycle.

Pay once on the first run; replay is free.