browser automation for AI agents
A Python harness that connects an LLM directly to a real browser via an editable CDP websocket, writing helper functions on the fly to self-heal.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
Use this when you want the mainstream Python browser-use library with Playwright and a rich agent API instead of raw CDP.
🌐 Make websites accessible for AI agents. Automate tasks online with ease.
Choose this if you need a TypeScript toolkit for browser-based AI agents.
Let AI agents browse the web. An autonomous toolkit for browser-based AI agents.
Choose this self-hostable JavaScript library when you want AI-agent browser automation in Node.js.
Open source and self-hostable browser automation library for AI agents
Choose this CDP powered browser agent if you want a minimal, similar approach with its own API.
Web-Use is a CDP powered Browser Agent
Choose this if you want browser infrastructure focused on cutting token spend for agents.
The Browser Infra that learns and cuts token spend by upto 90%
Choose this Python framework using direct CDP and CodeAgent if you want a similar architecture with a different API.
OpenBrowser is a framework for intelligent browser automation. It combines direct CDP communication with a CodeAgent architecture, where the LLM writes Python code executed in a persistent namespace, to navigate, interact with, and extract information from web pages autonomously.
Does the same job, but ships as an app. Useful to a person, not swappable into a codebase.
A Rust CLI tool, not an embeddable library; use it as a command-line utility or subprocess.
Browser automation CLI for AI agents
An MCP server exposing CDP browser actions; use it as a service, not code to embed.
让 Agent 直接操作真实 Chrome 的 MCP 服务，支持页面扫描、CDP、截图与物理输入
A CLI app that turns websites into command-line interfaces; an end-user tool.
Make Any Website into CLI & Use your logged-in browser by AI agent.
An MCP-native browser agent app with human-in-the-loop; a tool rather than a library.
Give your AI agent a real browser — with a human in the loop. Open-source MCP-native browser agent.
A Chrome extension app for AI web automation; an end-user product.
Open-Source Chrome extension for AI-powered web automation. Run multi-agent workflows using your own LLM API key. Alternative to OpenAI Operator.
A Rust CLI for CDP browser control; use it from the command line.
A tiny, fast Rust CLI that drives a real browser over the Chrome DevTools Protocol — built for coding agents.
A Rust CLI for browser control and CDP tasks; an end-user tool.
使用 agent-browser-cli 进行浏览器感知与控制。适用于标签页扫描/切换、页面 JS 执行、Cookie、CDP、contentSettings、截图、文件上传、下拉框点击、tmwd_cdp_bridge 初始化和 Web 工具排障
A Chrome extension GUI for DeepSeek Harness to control the browser; an end-user app.
Chrome sidebar extension that lets DeepSeek Harness operate your browser directly, no vision capabilities required. 一款 Chrome 侧边栏扩展程序，可让 DeepSeek Harness 直接操控您的浏览器，无需视觉能力。
An MCP server bridging CDP to clients; a service/tool, not a library.
MCP server that bridges clients to a real browser through CDP and a companion extension.
These projects were analysed and named browser-harness among their alternatives. The relationship is not symmetric — how browser-harness rates them is a separate judgement, made when browser-harness is analysed in its own right.
calls browser-harness “Self-healing Python harness for LLM browser tasks from the browser-use team; a peer library with enhanced stability.”
Let AI agents browse the web. An autonomous toolkit for browser-based AI agents.
calls browser-harness “Use if you want a Python harness for LLM browser tasks, similar to Stagehand but in Python.”
The SDK For Browser Agents
calls browser-harness “Choose this when you want a self-healing harness layer that helps LLMs complete browser tasks rather than manually orchestrating browser sessions on Steel.”
🔥 Open Source Browser API for AI Agents & Apps. Steel Browser is a batteries-included browser sandbox that lets you automate the web without worrying about infrastructure.
calls browser-harness “Choose if you want a higher-level agent harness rather than a raw MCP bridge to an existing browser session.”
MCP server that bridges clients to a real browser through CDP and a companion extension.