Dev Browser FAQs

Question 1

What is Dev Browser and what problem does it solve?

Accepted Answer

Dev Browser is an MCP server that enables OpenCode clients to automate local Chrome/Chromium tabs. It solves the need for AI agents to interact (click, type, navigate, screenshot) with existing local browser sessions without requiring Chrome's --remote-debugging-port flag.

Question 2

How does Dev Browser automate Chrome tabs without '--remote-debugging-port'?

Accepted Answer

It leverages a lightweight browser extension and a local relay. The extension connects to the relay, which then forwards Chrome DevTools Protocol (CDP) commands from OpenCode (via the MCP server) to the browser tab, bypassing the remote debugging port requirement.

Question 3

What are AI-friendly DOM snapshots and how are 'snapshotRef' identifiers used?

Accepted Answer

Dev Browser generates AI-friendly DOM snapshots in a YAML-ish outline, which include unique 'snapshotRef' identifiers. These 'snapshotRef's allow AI agents to reliably target and interact with specific elements (e.g., clicking or typing) based on the snapshot output.

Question 4

Can I control multiple browser tabs simultaneously with Dev Browser?

Accepted Answer

Yes, Dev Browser supports named, persistent tabs. You can open or reuse named tabs, allowing different OpenCode instances or agents to reliably control separate browser sessions without collisions, ensuring deterministic tab targeting for actions.

Question 5

What types of browser interactions can Dev Browser automate?

Accepted Answer

Dev Browser supports a wide range of browser interactions, including navigating to URLs, clicking elements by CSS selector or 'snapshotRef', typing into inputs, evaluating JavaScript, waiting for specific selectors, and capturing full-page or partial screenshots.

Dev Browser

Dev Browser

주요 기능

사용 사례

주요 기능

사용 사례