# agent-browser — AI Agent Browser Automation CLI (2026)

> agent-browser is a Rust CLI for AI agent browser automation; verified 32,919★ with snapshot refs, click/fill commands, and Chrome-for-Testing install.

## Install

Copy the content below into your project:

## Quick Use

```bash
npm install -g agent-browser
agent-browser install
agent-browser open https://example.com
agent-browser snapshot
agent-browser screenshot page.png
```

## Intro

agent-browser is a Rust CLI for AI agent browser automation; verified 32,919★ with snapshot refs, click/fill commands, and Chrome-for-Testing install.

**Best for:** Agent builders who want a fast, scriptable browser CLI without wiring Playwright into every project

**Works with:** macOS/Linux/Windows, Chrome (or Chrome for Testing), and CI scripts

**Setup time:** 2-5 minutes

## Main

- The core ergonomic win is `snapshot` refs (like `@e2`) so agents can click/fill by stable accessibility-derived handles instead of brittle CSS selectors.
- Installation stays pragmatic: npm/Homebrew/Cargo paths are documented, and `agent-browser install` can fetch Chrome from the official Chrome-for-Testing channel when needed.
- The command surface is intentionally CLI-first (open/click/fill/find/screenshot/pdf/stream), which makes it easy to wrap into scripts, task runners, or agent toolchains.
- For traditional automation, it still supports CSS selectors and role-based find commands; you can start with refs and drop to selectors when you need precision.

### FAQ

- **Do I need Playwright in my project?**: Not necessarily. The README positions agent-browser as a standalone CLI; you install it and drive Chrome via the CLI commands and refs.
- **Why run `agent-browser install`?**: It can download Chrome from the official Chrome for Testing channel (first run) and helps standardize automation environments across machines.
- **What should I verify first?**: Open a URL, run `snapshot`, then `click @e...` and take a screenshot so you validate the ref-based interaction loop end-to-end.

## Source & Thanks

> Source: https://github.com/vercel-labs/agent-browser
> License: Apache-2.0
> GitHub stars: 32921 · forks: 2035

---

<!-- ZH -->

## Quick Use

```bash
npm install -g agent-browser
agent-browser install
agent-browser open https://example.com
agent-browser snapshot
agent-browser screenshot page.png
```

## Intro

agent-browser 是 Rust 原生的 AI 代理浏览器自动化 CLI；已核验 32,919★，支持 snapshot 引用、click/fill 命令与 Chrome-for-Testing 安装。

**Best for:** 希望用高性能 CLI 做浏览器自动化、且不想在每个项目里集成 Playwright 的 agent 构建者

**Works with:** macOS/Linux/Windows、Chrome（或 Chrome for Testing）与 CI 脚本

**Setup time:** 2-5 minutes

## Main

- `snapshot` 输出带引用（如 `@e2`），让 agent 能基于可访问性树做 click/fill，避免依赖脆弱的 CSS 选择器，这是它最“agent-friendly”的点。
- 安装路径足够务实：npm/Homebrew/Cargo 都有说明，并且 `agent-browser install` 可按需从官方 Chrome-for-Testing 渠道拉取浏览器。
- 它把能力做成 CLI 命令面（open/click/fill/find/screenshot/pdf/stream），更适合作为脚本/任务编排/agent 工具链的一环。
- 需要更精细控制时仍支持传统选择器与基于 role 的查找，你可以先用引用跑通，再在关键步骤切换到 selector。

### FAQ

- **一定要在项目里装 Playwright 吗？**：不一定。README 把它定位成独立 CLI：安装后通过命令与引用驱动浏览器即可，无需把 Playwright 作为每个项目的依赖。
- **为什么要先跑 `agent-browser install`？**：它会在首次使用时按需下载 Chrome for Testing，并帮助在不同机器/CI 上统一可复现的自动化环境。
- **第一步怎么验证最靠谱？**：先 open 一个 URL，再 `snapshot`，然后用 `click @e...` 走通一次交互并截图，确认引用式控制闭环。

## Source & Thanks

> Source: https://github.com/vercel-labs/agent-browser
> License: Apache-2.0
> GitHub stars: 32921 · forks: 2035


---
Source: https://tokrepo.com/en/workflows/agent-browser-ai-agent-browser-automation-cli-2026
Author: Script Depot