browser-automation

Browser Automation

Use the browser tools when the task requires rendered state, authentication, client-side behavior, layout, accessibility, console output, or network traffic.

OpenCode normally exposes two compact tools:

• browser_help lists action names or loads the current schema for one action.
• browser_dispatch executes an action with { action, arguments }.

For compact browser_dispatch, put every action-specific field inside arguments, not beside it. For example, use { "action": "navigate_page", "arguments": { "url": "https://example.com" } }.

Common actions include navigate_page, take_snapshot, click, fill_form, take_screenshot, evaluate_script, list_console_messages, and list_network_requests. Call browser_help for an action before using it when its exact arguments are not already clear. Do not guess argument shapes.

Direct browser_* action tools may appear when compact tools are disabled. Use them directly with the same workflow.

Workflow

1. Reuse the current page by default. Use navigate_page to change URLs or reload after code or deployment changes. Use new_page only when simultaneous pages are necessary (max 5 owned pages per session; extra new_page calls fail, so close_page a finished page first), and close temporary pages when finished.
2. Navigate, then inspect the page before interacting.
3. Use the latest snapshot to identify elements and references. Refresh it after navigation or meaningful DOM changes.
4. Perform the smallest interaction needed, then inspect the result.
5. Verify consequential actions from visible state or relevant diagnostics.

Use screenshots when visual appearance, responsive layout, rendering, or image output matters. Prefer snapshots for locating elements and understanding page structure.

Inspect console messages for client-side exceptions and warnings. Inspect network requests when diagnosing API failures, redirects, authentication, caching, or unexpected response data.