Launching today

Browser for AI Agent

Launching today

Connect your browser and AI agents—in both directions

1 follower

Connect your browser and AI agents—in both directions

1 follower

Visit website

Browser for AI Agent is an open-source bridge between your browser and AI coding agents. Let any MCP- or CLI-capable agent read tabs, control the browser, run page tools, inspect errors, and take screenshots—all through a local native host. Or bring Claude Code, Codex, Cursor, and pi directly into DevTools or the browser side panel via ACP. Your browser data stays local and is shared only with the agents you configure.

Free

Launch tags:Browser Extensions•Chrome Extensions•Developer Tools

Launch Team

Wispr Flow: Dictation That Works EverywhereStop typing. Start speaking. 4x faster.

Promoted

NESBox

Maker

📌

Hey Product Hunt! 👋 I built Browser for AI Agent because browser automation for coding agents felt fragmented: one set of tools to control pages, another UI to talk with the agent, and too much cloud plumbing around sensitive browser data. This project connects both directions. Any MCP- or CLI-capable agent can read tabs, manage windows, run scripts or page tools, inspect errors, and take screenshots. In the other direction, Claude Code, Codex, Cursor, and pi can live inside DevTools or a browser side panel through ACP. It runs through a local native host, is open source, and supports Chrome, Edge, and Firefox. I’d love to hear what workflows you want next—and especially where the setup or permission model could be clearer.

Report

5d ago