AIエージェント
playwriter
Playwriter lets AI agents drive your real Chrome — your logins, extensions, and cookies — via Playwright
★ 4.0k スターMIT
ガイドを見る →BrowserOSは、browseros-aiが公開しているオープンソースのAIエージェントプロジェクトです（GitHub: browseros-ai/BrowserOS）。GitHubスターは13,844、フォークは1,479で、AGPL-3.0ライセンスの下で公開されています。GitHubでの最終更新は2026年10月7日です。
Open-source agentic browser your AI agents drive with your logins, plus an AI browser for humans.
Download it for your platform or install with Homebrew on macOS; it sits next to Chrome rather than replacing it.
One click brings over your logins, bookmarks, passwords and extensions so agents can work with your real accounts from the first task.
BrowserOS finds Claude Code, Codex, Cursor, VS Code and other MCP tools on your machine and connects any of them with one click.
Give the task to your agent (not the browser) and watch it work live in its own tab, or replay the saved session video later.
BrowserOS neo is a secondary browser made for AI agents, so agent work never tangles with your own browsing.
One-click Chrome import gives agents your logged-in accounts and persistent sessions - unlike headless drivers that start from a blank sandbox.
Several agents work at once in their own tabs, and the new-tab dashboard shows every agent's site, action and progress.
Every session is saved on your disk as a scrubbable video with a step-by-step action timeline, so you can rewind and audit exactly what happened.
Sessions, screenshots and history live under ~/.browserclaw/ and never leave your machine; only anonymous product-usage events are sent, with a one-toggle opt-out.
Any MCP-speaking tool connects - Claude Code, Codex, Cursor, VS Code, Zed, OpenCode and others - instead of locking you into one vendor's AI.
Clone the BrowserOS monorepo if you want to build or contribute; most users install the released browser instead.
git clone https://github.com/browseros-ai/BrowserOS.git
cd BrowserOSInstall the agent browser with Homebrew, or download installers for macOS, Windows or Linux from the project site.
brew tap browseros-ai/tap
brew install --cask browseros-neoImport your Chrome data, one-click connect Claude Code or Codex, then give the agent a web task from your usual tool.
# Connect your agent in BrowserOS neo - see README for the exact stepsBrowserOS is built by browseros-ai (Felafax, Inc.) with open-source contributors.
Both browsers are open source under AGPL-3.0 and free to download and use.
BrowserOS neo drives the agents you already use and pay for (such as a Claude Code subscription). The human BrowserOS browser lets you bring your own AI keys or run locally with Ollama.
Playwright-style drivers start a fresh, logged-out browser (good for CI, poor for real account work) and cloud browsers run on datacenter IPs that sites often block; BrowserOS runs locally with your real logins.
Install BrowserOS neo, import from Chrome in one click, connect your agent, and delegate your first task.
Playwriter lets AI agents drive your real Chrome — your logins, extensions, and cookies — via Playwright
Make your agents learn from experience with a dedicated context engine.
180 free gradients for CSS, Figma, and Sketch, plus a dataset for AI agents.
A batteries-included browser API so agents can automate the web.
An open benchmark for browser AI agents on everyday tasks.
A fast browser-automation CLI built for AI agents.
LibreAIカタログに掲載されたAIエージェントプロジェクトの中で最もスターの多いもの — BrowserOSが合わなかった場合に検討したい、現実的な候補リストです。
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
カタログの中でBrowserOSとスター数が近いAIエージェントプロジェクト — 前後の棚もそのまま見て回れます。
A framework for building, orchestrating and deploying AI agents and multi-agent workflows with support for Python and .NET.
Production-ready headless browser for AI agents, built in Rust - structure first, pixels only on demand.
Physical Intelligence's open vision-language-action models (pi0, pi0-FAST, pi0.5) for robot policies.
The official TypeScript SDK for Model Context Protocol servers and clients.
Secure sandboxes where AI agents run real code.
One portable, local-first memory layer for every AI agent - Markdown-native, user-owned and self-evolving.