Browser Use
LLM-driven, stealth-enabled browser automation with cloud scaling
A Python library and CLI that lets AI agents control browsers locally or via a stealth cloud, with templates, sandbox support, and easy authentication.
Overview
Browser-Use provides a Python library and command-line tool that lets large language models drive Chromium browsers. The Agent class pairs a task with the ChatBrowserUse model, handling navigation, clicks, typing, and data extraction. You can run browsers locally or switch to a stealth-enabled cloud instance to avoid detection and CAPTCHAs.
Getting Started & Production
Install with uv, add the browser-use package, and obtain a free API key for the cloud service. Use the sandbox decorator to run the agent next to the browser for minimal latency, or launch the persistent CLI for interactive debugging. Templates (default, advanced, tools) generate ready-to-run scripts, and the library includes helpers for custom tools, authentication profiles, and session persistence. For production, the Browser Use Cloud API scales browsers and manages memory, letting you focus on task logic.
Highlights
Pros
- Easy Python setup with uv and one-line install
- Flexible deployment: local, cloud, or sandbox
- Built-in support for custom tools and authentication profiles
- Comprehensive CLI for interactive debugging
Considerations
- Requires an LLM provider; token costs apply
- Cloud stealth browsers need an API key and internet access
- Chrome instances are memory-heavy, limiting parallelism locally
- Advanced configuration may need Python ≥3.11 and async familiarity
Fit guide
Great for
- Developers building AI assistants that need web interaction
- QA engineers automating form filling and data extraction
- Researchers prototyping browser-based LLM workflows
- Small teams needing quick, reproducible automation scripts
Not ideal when
- Environments with strict offline or no-internet policies
- Projects that cannot allocate Chrome's memory footprint
- Users seeking a pure no-code UI solution
- Scenarios requiring sub-second latency at massive scale without cloud
How teams use it
Job application auto-fill
Populates online application forms with resume data, reducing manual entry time
Instacart grocery ordering
Adds a predefined shopping list to the cart and proceeds to checkout
Custom PC parts research
Collects specifications and prices from multiple retailer sites into a summary report
Headline scraping for Hacker News
Retrieves top posts and stores titles for further analysis
Tech snapshot
Frequently asked questions
What's the best model to use?
ChatBrowserUse() is optimized for browser automation, delivering 3-5x faster completion with state-of-the-art accuracy.
Can I use custom tools with the agent?
Yes. Define functions with the `Tools` decorator and pass the instance to the `Agent` to extend its capabilities.
Is Browser-Use free to use?
The library is open source and free. You only pay for the LLM provider and optional cloud browser usage.
How do I handle authentication?
Reuse existing Chrome profiles, use temporary accounts via AgentMail, or sync profiles with the remote browser using the provided script.
How can I solve CAPTCHAs?
Use the stealth browsers on Browser Use Cloud, which include better fingerprinting and proxy support to avoid detection.
Project at a glance
Active- Stars
- 79,726
- Watchers
- 79,726
- Forks
- 9,423
Last synced 2 days ago