vercel-labs/agent-browser▌

Fast, persistent browser automation with session continuity across sequential agent commands. \n \n Supports three browser modes: headless Chromium, real Chrome with profile support, and cloud-hosted remote browsers with proxy configuration \n Includes 15+ command categories covering navigation, page inspection, interactions, data extraction, cookie management, and JavaScript execution \n Offers cloud session management, local server tunneling via Cloudflare, and parallel subagent execution thro

Automate any Electron desktop app via Chrome DevTools Protocol connection to agent-browser. \n \n Launch Electron apps with --remote-debugging-port flag, then connect agent-browser to that port for full snapshot-and-interact automation \n Works with any Electron app (Slack, VS Code, Discord, Figma, Notion, Spotify, etc.) across macOS, Linux, and Windows \n Supports multiple windows and webviews through tab switching; use named sessions to control multiple apps simultaneously \n Includes form fil

Headless Chrome browser automation inside ephemeral Vercel Sandbox microVMs for any Vercel-deployed framework. \n \n Spins up isolated Linux VMs on demand to run agent-browser commands, then shuts down automatically; works with Next.js, SvelteKit, Nuxt, Remix, Astro, and other Vercel frameworks \n Supports multi-step workflows with persistent browser sessions across sequential commands (navigation, form filling, screenshots, accessibility snapshots) \n Includes sandbox snapshots for sub-second s

$23

Systematically explore web applications to find bugs, UX issues, and quality problems with full reproduction evidence. \n \n Produces structured reports with step-by-step screenshots, repro videos, and detailed reproduction steps for every issue discovered \n Supports authenticated testing workflows, including OTP/email code entry, with session state persistence \n Distinguishes between interactive issues (requiring video and multi-step screenshots) and static issues (typos, visual glitches, req

Browser automation for checking messages, navigating channels, and extracting data from Slack workspaces. \n \n Connect to existing Slack sessions or open Slack in a browser, then use snapshots to identify interactive elements and take actions like clicking channels, searching, or reading messages \n Core workflow: connect, snapshot to find element refs, navigate to channels or tabs, extract text or capture screenshots as evidence \n Supports common tasks including checking unreads via Activity