Three things, done patiently.
No menus. No app to learn. Just a voice in the room, when the room is quiet.
Listens
Voice-first, hands-free, 24/7. No wake-word gymnastics — Oldaee is there when you start speaking, and it never rolls its eyes at the fifteenth telling of the same story.
Speech is captured locally and transcribed on-device. There is no always-on stream to a remote server.
Remembers
Names, faces in family stories, the way you take your tea. Oldaee builds a quiet memory of the household, so a question asked yesterday doesn't have to be re-explained today.
And when the same question comes back tomorrow — because sometimes it does — it answers again, with the same warmth, no impatience.
Responds
Calm, warm, in the user's own language. Sentences are short. Pace is slow. The voice is one chosen for clarity over personality.
Oldaee is calibrated for elderly cognition — gentler vocabulary, longer pauses, no jargon — without ever being patronising.
Privacy by architecture.
Oldaee runs on hardware in the home. Conversations stay there.
The voice models, the memory store, the language model — all of it is local. There's no account, no telemetry of conversation content, and no cloud transcript to leak. We chose this design because the people Oldaee is for can't always advocate for their own privacy. So we did it for them.
Stay in the loop
Drop your email and we'll write you when Oldaee is ready.
Thanks — we'll be in touch when Oldaee launches.