Story Kiosk
1 kiosk · wooden pieces, no touchscreen
You set a book down and pick a narrator. Gemini reads a passage you already know — then you swap a piece, and hear the same lines read a completely different way.
Google × Handstand
One floor, one evening. Each installation puts the Gemini Live audio models in front of a real person doing a real thing — ordering dinner, taking a swing, telling a story out loud.
24 September 2026 — days out
checking…
1 kiosk · wooden pieces, no touchscreen
You set a book down and pick a narrator. Gemini reads a passage you already know — then you swap a piece, and hear the same lines read a completely different way.
3 kiosks · 2 iPads each
You speak no Japanese. The chef speaks nothing else. You talk anyway — no headphones, no buttons — and the screens between you quietly assemble the order you agreed on.
2 bays · 2 cameras each
Two synchronized cameras measure the swing; Gemini interprets it and coaches you aloud, driving the replay screen to prove what it is saying. Interrupt it and ask.
The working sketches the demos were built out of — including the Phone Booth, which came off the floor on 2026-09-03 and lives on here — plus the station probes. Each one opens a live Gemini session on our own API key, so the lab asks for a password. Ask Sam or Glenn.
The floor's own instruments. Separate service, separate password.
/healthz from the platform service.