Technology
Two systems, designed so one doesn’t depend on the other.
The part that tells you where things are is designed to run on the glasses alone, with no phone and no internet. The camera is a separate part, used only when you ask for a description.
The sensing layer
Spatii is a pair of glasses. The electronics sit in the frame and temples, and it is worn the way ordinary glasses are worn.
Senses the space
A depth sensor reads the area in front of the wearer many times a second, building a picture of what’s near and what’s open, including objects at head and torso height that a cane can’t reach.
Conveys it as 3D audio you feel
Distance and direction are encoded as low-frequency 3D audio through bone-conduction transducers resting at the temples and behind the ears. A low obstacle is signalled behind the ear, one at head height at the temple, the side it comes from tells you left or right, and the closer it gets, the faster the cue. Turning your head is designed to show where an object starts and ends.
Bone conduction leaves the ears uncovered for speech, traffic and the other sounds people navigate by. Whether the cue ever gets in their way is one of the things the planned study will measure.
This layer is designed to run entirely on the device itself, without a network connection and without needing the phone to be awake or present.
The prototype
What we have actually built.
This is the current working prototype, revision 4: the hardware we build and test on today.
The whole assembly
A spectacle frame carrying the sensing electronics in the brow and both temples. Everything the device needs to work is on board.
The depth sensor
Set into the brow, above the bridge. It reads the space in front of the wearer continuously, building the picture of what is near and what is open.
The transducers
Bone-conduction transducers sit on each side, at the temple and behind the ear. They deliver the low-frequency cue that encodes distance and direction.
The camera
Mounted at the outer edge of the frame front. It is used only when the wearer asks for a description, and the sensing layer does not depend on it.
Buttons and charging
Physical controls along the top of each temple, reachable without looking, and a charge port on the left arm. No screen, and no phone required to operate the sensing layer.
The description layer
A camera and on-device artificial intelligence can describe a scene, recognise objects and help locate belongings, on request, when the wearer asks for it.
This layer is additive by design. The sensing layer does not depend on it, so if the description layer is switched off or unavailable, the sensing layer is intended to continue operating independently.
Private by default. The sensing layer runs on the device, and its data never leaves the glasses. The description layer operates only on explicit request, never continuously and never in the background.
Designed to sit alongside the cane, not replace it
The cane and the guide dog are proven mobility tools with decades of training practice behind them. Spatii is designed to add a layer of spatial awareness on top of them: a supplementary aid, used with the primary aid, never instead of it.
We’re deliberate about this language because we’re deliberate about the engineering behind it. A device that invites over-reliance is a device designed badly.
Want a say in how it works, or to try it when sessions open? Join the founding group. It takes a minute, and there’s nothing to buy or sign.