Browser Use

Browser Use

Browser Use is the leading open-source library for giving AI agents full browser control. Agents can navigate to URLs, click elements, fill forms, extract text, take screenshots, and handle multi-tab workflows — all driven by natural language tasks. It sits on top of Playwright and wraps browser state into a compact LLM-friendly representation.

Key Features

• Natural language tasks — describe a web task in plain English; the agent figures out the clicks
• Playwright-backed — full Chromium control with support for authentication, cookies, and JS-rendered pages
• Smart DOM extraction — converts complex page DOM into a compact, token-efficient representation for the LLM
• Multi-tab management — agents can open, switch, and close tabs as needed
• Vision + DOM — supports both screenshot-based and DOM-based navigation strategies
• Multi-agent — multiple Browser Use agents can collaborate in parallel
• Provider-agnostic — works with OpenAI, Anthropic, Gemini, and any LangChain-compatible model

Quick Start

pip install browser-use
playwright install chromium

import asyncio
from langchain_openai import ChatOpenAI
from browser_use import Agent

async def main():
    agent = Agent(
        task="Go to Hacker News and find today's top story title.",
        llm=ChatOpenAI(model="gpt-4o"),
    )
    result = await agent.run()
    print(result)

asyncio.run(main())

Install via ai-supply

npx ai-supply add browser-use-web-agent

Curated mirror of the open-source Browser Use project (MIT). Install upstream from the repository.