Ctrl+Z for reality.
Analyze a physical scene, then talk to REWIND directly. REWIND Voice uses the same deterministic SAVE → DIFF → REWIND → VERIFY engine as the external Alexa integration.
Image + REWIND Voice
Photo
Analyze this image with Nova. The raw image is not stored.
REWIND Voice
Microphone recognition depends on browser support. Typed commands always use the same REWIND Voice backend.
Privacy: the uploaded JPEG is sent to Amazon Nova for interpretation and discarded by this Vercel request. REWIND persists semantic physical state in DynamoDB, not the raw photo.
Ring Developer Playground
The existing REWIND Ring console connects to the Ring Developer Playground, receives live WHEP video, captures a real frame, and sends that frame through Nova into the same Physical State Protocol used by REWIND.
Start the existing Ring console from the repository before recording. Ring credentials and upstream WHEP session URLs stay server-side in the local process; they are never embedded in this public page.
- Run the secure Ring preview locally with the existing Ring environment already loaded.
- Open this mode in the public REWIND UI during the video so the evidence boundary is explicit.
- Open the live Ring console, select the Playground device, and click Start live view.
- Show the moving Ring Playground video, capture a frame, then click Observe with Nova.
- Hold on the validated semantic output long enough for judges to see Ring → WHEP → Nova → PSP.