SeeToSee · FJN Paper + AI communication

AI Communication Dashboard

One index for the full experiment. The current build turns the original capture bridge into a four-window desk: automatic context runs quietly, and your click on a live preview becomes the AI's intentional look here marker.

Open current v1.7 desk Read how v1.7 works Open original INS!EXPORT example

Current

The version to keep building

v1.7.0 CURRENT

SeeToSee AI Communication Desk

Designed around the workflow that proved useful during SeeToSee maintenance: multiple live browser views, short automatic context, deliberate screenshot markers, and one evidence ZIP for the AI.

4 windowsFour independently approved browser/window/screen streams.

5-sec contextBackground frames keep the sequence without a video.

Click = markerClick a preview to make a high-priority PNG.

INS!EXPORTMarkers + context + manifest in one AI handoff ZIP.

Why this is simpler

The old versions made the source, recorder, gallery, and AI destination feel like separate jobs. v1.7 keeps those capabilities but makes the live windows the center of the interface.

Launch it Version details

How the idea evolved

Every earlier build is preserved here so nothing useful was thrown away.

v1.6.0

INS!EXPORT Evidence BridgeAdded original full-resolution frames and AI-READ-FIRST evidence rules.

Open

v1.5.0

INS!EXPORT BridgeLocked the handoff to one predictable INS!EXPORT.zip for the current AI chat.

Open

v1.4.0

Three-slot ZIP HandoffMade bridge, approved source, and ChatGPT destination explicit.

Open

v1.3.0

ZIP Handoff BridgeCombined selected evidence into one ZIP instead of loose files.

Open

v1.2.0

Official Capture BridgeSeparated live viewing from saved official Steps and gallery.

Open

v1.1.0

Two-Tab AI BridgeIntroduced five-second frames, visible-change context, and manual shots.

Open

v1.0.0

1-on-1 AI BoothOriginal user-view + AI concept, local model experiment, and full-view control.

Open

What the ZIP actually contains now: v1.7 does not throw away the background sequence and it does not send only your clicked shots. Saved Steps contain both. The clicked marker frames are simply labeled as the highest-priority evidence, so the AI knows exactly where you were pointing.