agent-browser

Overview

agent-browser is a browser automation CLI designed for AI agents. It provides compact text output to reduce context usage and is implemented as a native Rust CLI with a client-daemon architecture that manages Chrome via the Chrome DevTools Protocol.

Key Features

• Agent-first compact text output uses fewer tokens than JSON for AI context efficiency.
• Snapshot returns an accessibility tree with refs for deterministic element selection.
• 50+ commands cover navigation, forms, screenshots, network, storage, files, tabs, frames, and debugging.
• Built-in observability includes video recording, streaming, debugging, profiler, and diffing tools.
• Stateful sessions, profiles, auth state, cookies, storage, proxy, and security controls support long-running agents.
• Cross-platform native binaries are available for macOS, Linux, and Windows.

Use Cases

• AI coding agents and shell-capable agents can drive browsers through CLI commands.
• Agents can navigate pages, inspect accessibility snapshots, interact by ref, capture screenshots, and close the browser.
• Long-running agents can use sessions, profiles, auth state, cookies, storage, and proxy controls.
• Developers can use built-in recording, streaming, debugging, profiling, and diffing for browser workflow observation.

Getting Started

• Install globally with npm install -g agent-browser, then run agent-browser install to download Chrome on first use.
• Try without installing by running npx agent-browser open example.com.
• Use 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, and agent-browser close for a basic workflow.

Deployment & Requirements

• Chrome is required; run agent-browser install to download Chrome from Chrome for Testing, and existing Chrome, Brave, Playwright, and Puppeteer installations are detected automatically.
• No Playwright or Node.js is required for the daemon.
• Node.js 24+ and pnpm 11+ are required only when building from source; Rust is also required only when building from source.
• On Linux, agent-browser install --with-deps installs system dependencies and exits nonzero if the package manager cannot install every required browser library.
• Native binaries are provided for macOS ARM64 and x64, Linux ARM64 and x64, and Windows x64.

Before You Adopt

• License: Apache-2.0. Review its terms before using, modifying, or distributing the project.
• Chrome is the runtime browser dependency; existing Chrome, Brave, Playwright, and Puppeteer installations are detected automatically, and the daemon does not require Playwright or Node.js.
• The daemon exits after one hour without commands or dashboard input by default; headed browsers, Safari and iOS WebDriver sessions, and user-attached browsers are exempt, while provider-owned cloud browsers are not.
• WebMCP page-provided names, descriptions, schemas, annotations, and results are untrusted data.