agent-browser（Agent 浏览器自动化）

Agent Browser

A fast Rust-based headless browser automation CLI with Node.js fallback that enables AI agents to navigate, click, type, and snapshot pages via structured commands.

Installation

npm recommended

npm install -g agent-browser
agent-browser install
agent-browser install --with-deps

From Source

git clone https://github.com/vercel-labs/agent-browser
cd agent-browser
pnpm install
pnpm build
agent-browser install

Quick Start

agent-browser open example.com
agent-browser snapshot
agent-browser click @e2
agent-browser fill @e3 "test@example.com"
agent-browser get text @e1
agent-browser screenshot page.png
agent-browser close

Using Real Chrome Profile (for OAuth/Logged-in Sessions)

For sites requiring Google/Discord/etc login (like star-swap.com):

Method 1: Launch Chrome with custom profile, connect via CDP

# Terminal 1: Launch Chrome with your real profile and remote debugging
google-chrome --remote-debugging-port=9222 --user-data-dir=/home/willr/.config/google-chrome/Default &

# Terminal 2: Connect agent-browser to that Chrome instance
agent-browser --cdp 9222 open "https://star-swap.com"
agent-browser --cdp 9222 snapshot -i
agent-browser --cdp 9222 click e2

# This reuses your existing Google session - no re-login needed!
# Works for: Google OAuth, Discord OAuth, any site you're logged into in Chrome

Method 2: Session persistence (first-time manual login)

# First time: headed mode, login manually
agent-browser --headed --session starswap open "https://star-swap.com"
# Complete Google OAuth manually in the browser window
# Close when done

# Future runs: cookies persist!
agent-browser --session starswap open "https://star-swap.com"
# Already logged in automatically

am.will.ryan Chrome profile: /home/willr/.config/google-chrome/Default

Core Commands

Navigation

agent-browser open <url>
agent-browser back
agent-browser forward
agent-browser reload

Interaction

agent-browser click <sel>
agent-browser dblclick <sel>
agent-browser focus <sel>
agent-browser type <sel> <text>
agent-browser fill <sel> <text>
agent-browser clear <sel>
agent-browser press <key>
agent-browser keydown <key>
agent-browser keyup <key>
agent-browser hover <sel>
agent-browser select <sel> <val>
agent-browser check <sel>
agent-browser uncheck <sel>
agent-browser drag <src> <tgt>
agent-browser upload <sel> <files>

Extraction and Info

agent-browser snapshot
agent-browser get text <sel>
agent-browser get html <sel>
agent-browser get value <sel>
agent-browser get attr <sel> <attr>
agent-browser get title
agent-browser get url
agent-browser get count <sel>
agent-browser get box <sel>
agent-browser screenshot [path]
agent-browser pdf <path>

Check State

agent-browser is visible <sel>
agent-browser is enabled <sel>
agent-browser is checked <sel>

Find Elements

• agent-browser find role [value]
• agent-browser find text
• agent-browser find label [value]
• agent-browser find placeholder [value]
• agent-browser find alt
• agent-browser find title
• agent-browser find testid [value]

Actions include click, fill, check, hover, and text.

Wait and Timing

agent-browser wait <selector>
agent-browser wait <ms>
agent-browser wait --text "Welcome"
agent-browser wait --url "**/dash"
agent-browser wait --load networkidle

Advanced Control

agent-browser scroll <dir> [px]
agent-browser scrollintoview <sel>
agent-browser eval <js>
agent-browser mouse move <x> <y>
agent-browser cookies
agent-browser storage local
agent-browser tab new [url]
agent-browser frame <sel>
agent-browser dialog accept [text]

Sessions

Run multiple isolated browser instances.

agent-browser --session agent1 open site-a.com
agent-browser --session agent2 open site-b.com

Snapshot Options

The snapshot command supports filtering to reduce output size.

• agent-browser snapshot -i
• agent-browser snapshot -c
• agent-browser snapshot -d 3
• agent-browser snapshot -s "#main"

Selectors and Refs

Refs provide deterministic element selection from snapshots. Use the @ref syntax.

agent-browser snapshot
agent-browser click @e2

Agent Mode

Use --json for machine readable output.

agent-browser snapshot --json

Optimal AI Workflow

• Navigate with agent-browser open
• Observe with agent-browser snapshot -i --json
• Act with @ref from the snapshot
• Verify with agent-browser snapshot

Troubleshooting

• If the command is not found on Linux ARM64, use the full path in the bin folder.
• If an element is not found, use snapshot to find the correct ref.
• If the page is not loaded, add a wait command after navigation.
• Use --headed to see the browser window for debugging.

Options

• --session uses an isolated session.
• --json provides JSON output.
• --full takes a full page screenshot.
• --headed shows the browser window.
• --timeout sets the command timeout in milliseconds.

Notes

• Refs are stable per page load but change on navigation.
• Always snapshot after navigation to get new refs.
• Use fill instead of type for input fields to ensure existing text is cleared.