It doesn't just watch.
It talks back.
Ocula sees who's at your door, describes them out loud, and speaks in a real human voice. Not an alarm. A person, watching, who isn't afraid to say what they see.
FIRST UNITS SHIP TO TORONTO · NO SPAM, JUST LAUNCH DAY
Not sure it's for you? Take the 30-second quiz →You're on the list. We see you, the good kind.
> "Yeah, I see you. Wrong house."
By the time you watch the footage,
they're already inside.
A thief expects to be filmed. They don't expect the house to look right at them and describe what they're wearing. That's the part that makes someone leave.
It spots a person
On-device AI catches someone the second they enter frame, day or night. No cloud lag.
It describes them
The hoodie, the gloves, the hands in their pockets. Specific enough that they know they've been seen.
It talks back
A calm, human voice says exactly what it sees. That's what makes them turn around.
Ask a burglar what makes them leave. It's not the alarm, and it's not the camera. It's realizing someone inside already knows they're there.
▌ FROM A STUDY OF 422 CONVICTED BURGLARS · UNC CHARLOTTE, 2013
What are you actually worried about?
Five quick questions, about thirty seconds. We'll tell you straight, and you'll help decide what ships first.
What your camera and alarm can't do.
Questions, answered straight.
Is this real, or just a concept?+
Real, and running today. Ocula is a live pipeline, not a single gadget: on-device vision catches a person the instant they enter frame, a vision-language model reads the scene and writes a specific line about that exact person, and a natural-voice engine speaks it, start to finish in about four seconds. Current units run on compact edge hardware built into the camera; production ships on hardware purpose-built for the pipeline. The hard part isn't the box, it's getting a machine to see, judge, and speak like a person in real time. The waitlist is for that first run.
Does it actually sound human, or robotic?+
Human. It uses a real, natural-sounding voice, not a flat text-to-speech beep. On setup you can even have it speak in a voice you choose.
What does it cost?+
Pricing is being finalized for the first run. Join the waitlist and you'll get the launch price before anyone else, with no commitment now.
Do I need to be in Toronto?+
First units ship to Toronto, where break-ins jumped 552% in 2025. The waitlist is open everywhere, and the quiz helps us decide where to ship next.
What about my privacy?+
Detection runs on the device, not in someone else's cloud. We're building it privacy-first and will be upfront about exactly what's stored and what isn't.
Be the first one it speaks for.
BUILT IN TORONTO BY A 17-YEAR-OLD · LIMITED FIRST RUN
You're on the list. We see you, the good kind.