Browser Agent Recorder

Record browser workflows and export local SOPs and Agent Skill Packs.

Record a browser task once. Replay it forever. Browser Agent Recorder captures what you do in the browser — clicks, typing, navigations, dropdowns, file uploads, and dialogs — with a screenshot of every step that highlights exactly what you clicked. Then it turns that recording into documentation a human or an AI agent can follow. WHAT IT CAPTURES • Clicks, inputs (with IME/Chinese composition handling), form changes, submits • Page navigations and single-page-app route changes • Keyboard shortcuts, file uploads, drag & drop • Native alert / confirm / prompt dialogs • A highlighted screenshot per step EDIT & ORGANIZE • Rename and describe steps; drag to reorder; delete and restore • Set the workflow goal and success criteria • Insert manual note / wait steps • Edit selectors and see confidence scores EXPORT • Markdown SOP — a readable how-to guide with screenshots • Playwright test (.ts) — semantic locators • Chrome DevTools Recorder (.json) • Agent Skill Pack (.zip) — a full bundle for autonomous AI agents, with protected trajectory, append-only learning notes, auth/safety policies, and replay code PRIVATE BY DESIGN • 100% local. No backend, no login, no cloud sync, no analytics. • Passwords are never stored; sensitive fields are masked and redacted from screenshots. Press Alt+Shift+R to start or stop recording anywhere.