agent-browser
๐Pluginvercel-labs/agent-browser
Installation
/plugin marketplace add vercel-labs/agent-browser/plugin install agent-browser@agent-browser
|A browser automation CLI for AI agents by Vercel Labs, enabling website navigation, form filling, button clicking, screenshot capture, and data extraction through simple commands.
Overview
Agent Browser is a Claude Code skill from Vercel Labs that automates web browsing tasks. It navigates websites, extracts content, interacts with page elements, and captures screenshots using a programmatic browser automation workflow.
Key Features
- Navigate to URLs with
agent-browser open, take interactive snapshots with element references (@e1, @e2) - Interact with page elements using references to click, fill forms, and select options
- Capture full-page and element-specific screenshots for visual verification
- Supports complex multi-step browsing flows with automatic state management
Who is this for?
Developers who need to automate web interactions, scrape content, test web UIs, or capture screenshots as part of their Claude Code workflow.
Part of
vercel-labs-agent-browser
More from this repository10
Automates web browsing tasks by navigating websites, extracting content, interacting with elements, and capturing screenshots programmatically
Browser automation CLI for AI agents
Systematically explores and tests web applications to find bugs, UX issues, and other problems.
Skill for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Connects to running Electron apps and controls them programmatically.
Skill from agent-browser, a headless browser automation CLI for AI agents with a fast Rust binary and Node.js fallback for cross-platform support.
Skill
A concise, practical guide for creating specialized skills that extend Claude's capabilities with domain-specific knowledge and workflows.
Run agent-browser browser automation on AWS Bedrock AgentCore cloud browser sessions, resolving credentials automatically from environment variables or the AWS CLI. All standard agent-browser commands (open, snapshot, click, screenshot, close) work identically โ only the browser runtime changes to a managed AWS-hosted session.
Browser automation CLI for AI agents, built as a fast native Rust binary. Supports opening pages, taking snapshots, clicking, typing, and extracting content, with automatic detection of existing Chrome, Brave, Playwright, and Puppeteer installations.
Skill