
stagehand
AgentSDK for browser agents with Playwright-style API and self-healing actions
- Open source
- Agent framework
- 26.1K GitHub stars
- Browser automation

SDK for browser agents with Playwright-style API and self-healing actions

Browser engine and CLI built for agents that browse, extract and crawl at scale

Browser agent that picks from a page's real controls and cites its sources

Autonomous web testing agent that checks functionality, performance, UX and security

Open-source Chrome extension that automates the browser from plain-language instructions

Self-hosted browser agent that scrapes protected sites from a plain-English prompt
Build and continue browser and Mac automation scripts for Super through Claude
Cloud Android phones for agents to see, tap, type and install apps over MCP
Persistent Linux cloud desktops that an agent can see, click, type into and run tasks on
Not checked yetBulk website screenshots in desktop, tablet and mobile sizes, with cookie banners hidden
Not checked yetPay-per-request screenshots, page-to-Markdown or PDF and Nostr relay tests via Lightning
Not checked yetFull-page screenshots and URL-to-PDF in real Chrome, with cookie banners rejected
Not checked yetBulk website screenshots in full-page, mobile or element views as PNG, JPEG or PDF
Not checked yetChrome extension plus MCP server so coding agents can use your signed-in Chrome profiles
Not checked yetHeadless Metabase MCP server using an API key, with 187 tools and read-only mode
Runs locallyCapture full-page website screenshots and PDFs for lists of URLs through Apify
Not checked yetHosted cloud browsers for agents with saved logins, residential IPs and live takeover
Not checked yetSupervised cloud browser where an agent drives and a human can watch and take over
Not checked yetHosted browser sessions for agents over MCP or REST, paid per session in stablecoins
Not checked yetGives an AI agent your real Chrome through compact text snapshots and a visible cursor
Runs locally
Remote MCP server that runs verified actions on mapped websites instead of browser control
Not checked yetJob postings from Workday, Greenhouse, Lever and Ashby, with change tracking
Not checked yetCloud browser sessions for AI agents with live view and human handoff
Not checked yet
Full-page website screenshots and URL-to-PDF, with cookie banners removed
Not checked yetRemote MCP for web search, page extraction and controlled browser sessions via OAuth
Not checked yetWeb data extraction and browser workflow tools for AI assistants, scoped by API key
Not checked yetCapture website screenshots with local Chrome from a CLI or an AI agent
Not checked yetControl cloud Android phones on US mobile IPs: screenshots, taps, installs and runs
Not checked yetHosted MCP server for running cloud browsers and managing Kernel apps from an agent
Not checked yetRead-only website smoke tests in a real browser, run by your AI agent over MCP
Not checked yet