Browser Use

Browser Use is an open-source Python framework that makes websites accessible to AI. It is the most popular open-source project for building AI agents that operate a browser: instead of calling brittle REST APIs, an agent sees the page, reasons about it, and takes actions — clicking, typing, navigating, and extracting data just like a human user would. The library is trusted across the industry and has accumulated one of the largest followings of any agent project, with hundreds of thousands of GitHub stars. Under the hood, Browser Use is built on Playwright and wraps the browser in a thin, self-healing harness. The harness abstracts away the messy details of web interaction — dynamic content, popups, layout shifts — so the agent can focus on the task. The newest versions ship a beta agent powered by a Rust core, designed for speed and reliability with current frontier models, and the framework reports benchmark success rates approaching 90%. Developers choose Browser Use because it works with any LLM provider and is model-agnostic: you point it at OpenAI, Anthropic, Google, or local models, and equip the agent with custom tools. It supports persistent filesystem and memory across runs, and for production deployments it pairs with cloud browsers that add proxy rotation, captcha solving, and stealth — or you can self-host the open-source agent entirely on your own machines. Typical use cases include automating workflows that live behind logins or complex UIs, scraping data from sites without APIs, filling out multi-step forms, testing web applications, and building autonomous assistants that complete end-to-end web tasks. For anyone building a web-operating agent, Browser Use is the default starting point.