Browser Use

About

Browser Use is an LLM-powered browser automation tool that allows language models to control a real browser via structured system and user prompts. The LLM decides which browser actions to perform (open pages, navigate, click, read content), while the execution is handled by the browser runtime. The integration works with any supported LLM and does not rely on model-specific features.

Installation

Environment requirements

• Python 3.12

• Supported operating systems: macOS / Linux / Windows

• Chromium is used (installed automatically)

Environment setup

pip install uv
uv venv --python 3.12
source .venv/bin/activate
# Windows:
.venv\Scripts\activate

Installing dependencies

uv pip install browser-use python-dotenv
uvx browser-use install

How to Use AIML API with Browser Use

General concept

Browser Use works as follows:

• The user provides system and user prompts to the LLM.

• The LLM returns instructions.

• The browser executes the actions.

Important notes:

• It does not require any special tool-calling APIs.

• It does not rely on model-specific features.

• All decision-making logic lives entirely in the LLM.

• The browser only executes the commands returned by the model.

There is no tight coupling to a specific model. All our LLMs are supported.

---

Setting up your API key

You can create a dedicated .env file and store your AIMLAPI key there once, then reference this file from your code, as shown in the example code below.

AIMLAPI_KEY=YOUR_KEY

Alternatively, you can place the key directly in your code, as shown in all our model call examples in the API REFERENCES section of this portal.

---

Example Code

Python

import asyncio
import os
from dotenv import load_dotenv
from browser_use import Agent, ChatOpenAI

load_dotenv()

llm = ChatOpenAI(
    model = "x-ai/grok-4-07-09",             # Insert the ID of one of our LLMs
    base_url = "https://api.aimlapi.com",
    api_key = os.getenv("<YOUR_AIMLAPI_KEY>"),        # Insert your AIMLAPI Key
)

agent = Agent(
    task = 'Find the number of stars of the browser-use repo', # Insert the instruction
    llm = llm,
    use_vision = True,
)

async def main():
    await agent.run(max_steps=10)

asyncio.run(main())

Supported Models

All our LLMs are supported.

More

For further information about the framework, please check the official documentation.

For additional examples, check out the repo.