dom0

Browser automation for AI agents

dom0 enables programmatic browser automation through a local command-line interface. It connects to a locally-running CLI daemon via WebSocket on localhost to receive commands and return results. Features: - Capture accessibility tree snapshots for AI-readable page structure - Navigate, click, type, scroll, and interact with page elements - Take screenshots and export pages to PDF - Manage tabs and windows - Read and set cookies, localStorage, and sessionStorage - Emulate devices, viewports, and media preferences How it works: dom0 uses the Chrome DevTools Protocol (via the debugger API) to automate browser interactions. Commands are sent from the local dom0 CLI tool over a localhost-only WebSocket connection. No data is sent to any remote server. Permissions: - debugger: Required for Chrome DevTools Protocol access to automate page interactions - tabs: Required to list, create, switch, and close browser tabs - activeTab: Required to attach the debugger to the currently active tab - storage: Required to persist extension configuration - alarms: Required to keep the background service worker alive for the WebSocket connection - Host permissions (all URLs): Required to automate any website the user navigates to Privacy: All communication stays on your local machine. dom0 connects only to localhost and never transmits browsing data to external servers. See our privacy policy at https://bytespace.ai/privacy-policy. Learn more about dom0 at https://bytespace.ai or https://dom0.dev