Skip to main content
This guide is an end-to-end setup for running a browser-use agent with the Laminar Debugger so you can rerun from checkpoints.

​
1) Install prerequisites

Install uv if you do not already have it: 
pip install -U uv
Create a virtual environment and install dependencies: 
uv venv --python 3.11
source .venv/bin/activate
uv pip install lmnr browser-use python-dotenv
Install the Browser Use Chromium bundle: 
uvx browser-use install
If your Browser Use LLM provider needs extra packages or env vars, follow the provider setup in Browser Use docs.

​
2) Set environment variables

Create a .env file in your project root: 
LMNR_PROJECT_API_KEY=your_laminar_project_api_key
If you are using ChatBrowserUse (recommended), add: 
BROWSER_USE_API_KEY=your_browser_use_api_key
If you prefer Claude, add: 
ANTHROPIC_API_KEY=your_anthropic_api_key
CLAUDE_MODEL=your_claude_model_name

​
3) Create the entrypoint

import os
import asyncio
from dotenv import load_dotenv
from browser_use import Agent, ChatBrowserUse, ChatAnthropic
from lmnr import observe, Laminar

load_dotenv()
Laminar.initialize(project_api_key=os.getenv("LMNR_PROJECT_API_KEY"))

@observe(rollout_entrypoint=True)
async def main(prompt: str):
    provider = os.getenv("LLM_PROVIDER", "chatbrowseruse")
    if provider == "anthropic":
        llm = ChatAnthropic(model=os.getenv("CLAUDE_MODEL"), temperature=0.0)
    else:
        llm = ChatBrowserUse()

    agent = Agent(task=prompt, llm=llm)
    history = await agent.run()
    print(history.final_result())

if __name__ == "__main__":
    asyncio.run(
        main(
            "go to ycombinator.com, summarize 3 startups from the summer 2025 batch."
            # "go to laminar.sh, summarize the pricing page."
        )
    )
Notes:
  • The rollout_entrypoint=True flag marks this function for debugger sessions.
  • Keep the CLI running while you iterate; the debugger will rerun using the latest code on disk.

​
4) Start a debugger session

npx lmnr-cli@latest dev path/to/entrypoint.py
If the file exposes multiple debugger entrypoints, pass --function main.
[Screenshot: terminal output with debugger session link]

​
5) Open the debugger in Laminar

Open Debugger in your project, select the session, provide input arguments as JSON, and click Run. Debugger session view

​
6) Make a change and rerun (Debugger workflow)

Use the Debugger to iterate without restarting your local process.
  1. Open the trace in Tree or Reader view.
  2. Click the checkpoint icon on a span and choose Run from here.
  3. Expand System Prompts and toggle Override System Prompt on for the Browser Automation Agent.
System prompt override toggle
  1. Edit the prompt in place, then click Run again to rerun with the updated instructions.
Notes:
  • Keep the CLI running while you iterate; the debugger reruns using the latest code on disk.
  • If you changed dependencies, restart the CLI.