Record Actions.
Generate SOPs.
Browser action recording + voice narration → Structured SOPs for AI agents. Do it once, let AI do it a thousand times.
Built for the AI Native Era
Transforming how SOPs are created and presented, making them readable for both humans and AI.
Real-time Streaming STT
Enterprise-grade real-time speech transcription with multi-language auto-detection. Narrate while you operate; your voice is automatically linked to the corresponding actions without post-editing.
Smart Action Capture
Automatically records clicks, inputs, and scrolls. Precisely distinguishes between 'text selection' and 'clicks', intelligently filtering out meaningless redundant actions.
Annotated Screenshots
Auto-captures the screen on every action, marking the exact click position with a high-contrast ring, ensuring clarity on both light and dark backgrounds.
Standalone Structured Export
One-click export to a self-contained HTML file. Send it to LLMs (like Claude or GPT-4V), and the AI instantly understands your intent and the exact CSS Node path.
What can you build with it?
Teach AI tasks
Record a workflow once, export the structured SOP with selectors, and feed it to a Browser Agent for automated execution.
Product Manuals
Demonstrate and dictate simultaneously to directly generate visually rich user tutorial documents, eliminating tedious screenshot stitching.
Bug Reproduction
The ultimate weapon for QA. Every click path, page state, and operator's supplementary explanation are perfectly bound.
Internal Training
For employee handover; just record routine backend operations once, and newcomers can self-onboard following the steps.
Three Steps to Automation
1. Install
Download source code and 'Load unpacked' in Chrome. Open the sidebar.
2. Open Target Workflow
Navigate to the workflow page you want to demonstrate and keep the sidebar ready for recording.
3. Record
Click 'Start Recording', operate and narrate. Export HTML when done.