> ## Documentation Index
> Fetch the complete documentation index at: https://www.bugzy.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Browser automation

> Playwright-driven browser automation inside isolated Cloud Run containers — with screenshots, video recording, and trace files.

Bugzy uses Playwright to interact with your application the way a real user would — navigating pages, filling forms, clicking buttons, and verifying visual output. Browser automation runs inside isolated Cloud Run containers with pre-installed Chromium browsers.

## How it works

Every Bugzy execution launches a headless Chromium browser inside a sandboxed container. The browser-automation subagent controls this browser through the Playwright CLI, which provides direct page interaction without consuming agent context with tool schemas.

```
Cloud Run Container
├── Chromium (pre-installed, version-locked)
├── Playwright CLI (browser control)
└── Agent (issues navigation + interaction commands)
```

The agent can:

* Navigate to any page within your application's base URL
* Click buttons, fill forms, select dropdowns, and interact with dynamic UI
* Wait for elements, network requests, and page transitions
* Execute JavaScript in the page context
* Simulate mobile viewports and different screen sizes

## Evidence capture

Bugzy captures visual evidence during every interaction:

| Artifact         | When captured                             | Purpose                                            |
| ---------------- | ----------------------------------------- | -------------------------------------------------- |
| Screenshots      | On failure, and at key interaction points | Debugging, bug reports                             |
| Video recordings | Full test execution (when enabled)        | Reviewing test flow                                |
| Trace files      | On failure                                | Step-by-step replay with network and DOM snapshots |

Screenshots are automatically attached to bug reports filed in your issue tracker and to messages posted in Slack or Teams.

## Domain boundary

Browser automation is scoped to your project's configured base URL domain. Bugzy does not navigate to external sites or third-party services during test execution. This boundary is enforced at the task level to prevent unintended interactions with production services outside your application.

## Cross-browser support

The container ships with Chromium pre-installed and version-locked to match the Playwright version. Firefox and WebKit are available for cross-browser testing when configured in your test suite.

## Learn more

<CardGroup cols={2}>
  <Card title="Execution environment" icon="server" href="/docs/platform/execution-environment">
    Container isolation, the execution pipeline, and where test artifacts live.
  </Card>

  <Card title="Extend test coverage" icon="shield-plus" href="/docs/capabilities/extend-coverage">
    How browser exploration feeds into test plan generation.
  </Card>
</CardGroup>