BrowserOS logo

BrowserOS

AI‑powered Chromium browser that runs agents locally, privacy‑first

A Chromium‑based browser that lets you run AI agents directly on your machine using your own API keys or local models, keeping data private while supporting extensions and MCP integration.

BrowserOS banner

Overview

Overview

BrowserOS brings AI agents into a familiar Chromium browser, letting power users and developers run large language models directly on their machine. By using your own OpenAI, Anthropic, or local Ollama keys, all prompts and browsing history stay on‑device.

Core capabilities

The browser imports Chrome data, supports the full range of Chrome extensions, and connects to AI providers or locally hosted models. An optional MCP server lets external tools such as claude-code or gemini-cli drive the browser for automated testing or data scraping. Upcoming features include a built‑in AI ad blocker.

Getting started

Install the macOS dmg, Windows installer, AppImage, or Debian package, then configure your API key or local model. BrowserOS is released under AGPL‑3.0, with an active Discord and Slack community that welcomes contributions and feature ideas.

Highlights

Runs AI agents directly in the browser using local or user‑provided API keys
Full Chrome extension compatibility with a familiar UI
MCP server enables control from external tools like claude‑code or gemini‑cli
Privacy‑first design keeps browsing data on‑device

Pros

  • Local AI execution protects user privacy
  • Supports existing Chrome extensions
  • Cross‑platform installers for macOS, Windows, and Linux
  • Open‑source community driven development

Considerations

  • Requires setup of API keys or local models, adding initial complexity
  • May lack some proprietary Chrome features
  • AI ad blocker is still in development
  • Performance depends on local hardware and model size

Managed products teams compare with

When teams consider BrowserOS, these hosted platforms usually appear on the same shortlist.

Perplexity Comet logo

Perplexity Comet

AI-powered browser with conversational interface

Looking for a hosted option? These are the services engineering teams benchmark against before choosing open source.

Fit guide

Great for

  • Users who want AI‑assisted browsing without sending data to cloud services
  • Developers building or customizing AI agents in a browser environment
  • Privacy‑focused professionals handling sensitive information
  • Teams needing cross‑platform, extension‑compatible browsers with AI automation

Not ideal when

  • Casual users preferring a plug‑and‑play experience without configuring API keys
  • Environments that rely on Chrome’s proprietary services such as Google Sync
  • Low‑end devices unable to run local large language models efficiently
  • Organizations requiring built‑in ad‑blocking before the feature is released

How teams use it

Automate repetitive web tasks

AI agents fill forms, navigate sites, and extract data, freeing users from manual clicks.

Secure AI‑driven research

Run local LLMs to summarize articles while keeping browsing history on the device.

Integrate browser into development pipelines

Control BrowserOS via MCP from claude‑code to test UI flows programmatically.

Extend existing Chrome extensions with AI

Leverage familiar extensions while adding agentic capabilities without rewriting code.

Tech snapshot

C++54%
Python29%
TypeScript6%
HTML5%
Objective-C++3%
C3%

Tags

browserosmacosllmlmstudiohacktoberfestollamabrowseragentchromiumwindowslinux

Frequently asked questions

Do I need an internet connection to use BrowserOS?

You need internet to access web pages and to call remote AI services; local models via Ollama can run offline after they are downloaded.

Which AI providers are supported?

OpenAI, Anthropic, and any local model accessible through Ollama or LMStudio.

Is my browsing data sent to any third party?

No. All data stays on your device unless you explicitly send it to a remote API.

Can I use my existing Chrome extensions?

Yes. BrowserOS uses the same Chromium extension framework, so most extensions work out of the box.

What platforms are supported?

macOS, Windows, and Linux (AppImage and Debian packages).

Project at a glance

Active
Stars
8,732
Watchers
8,732
Forks
847
LicenseAGPL-3.0
Repo age8 months old
Last commit10 hours ago
Primary languageC++

Last synced 3 hours ago