Give the AI a small window into the work.
One tiny browser companion can share an approved tab, window or whole screen; measure and crop an area; mark it up; record it; save important moments; and package the evidence so an AI can understand what happened.
What do you want the AI to see?
Nothing is shared until you choose it.
Seven capabilities, one at a time.
The walkthrough is organized around jobs the assistant performs, not version numbers. The historical builds stay available in the library.
See
Approve a browser tab, app window, or entire screen through the browser’s own share picker.
Open →Measure & crop
Watch X/Y and width/height live, then fine-tune the box down to individual source pixels before saving anything.
Open →Precision Measure →
Mark & record
Pen, highlighter, redact, arrow, undo, clear, PNG capture, and cropped WebM recording.
Open →Remember steps
Capture context on a timer, mark important moments, and keep official steps in a gallery.
Open →Send evidence
Bundle the selected steps into one INS!EXPORT.zip with an AI-readable evidence contract.
Open →Watch the desk
Keep up to four separately approved browser/window/screen sources visible at once.
Open →Know the boundary
No silent tab discovery, hidden recording, raw cross-tab clicks, or automatic typing into another AI conversation.
Open →