01
Describe the surroundings
Ask "what's in front of me?" and get one clear spoken answer: object, distance, direction.
Prototype · Designed in Almaty
Assistance that stays with you.
OKO are offline-first AI glasses that describe what's in front of you: one short spoken sentence, no internet required for core use. We build in the open and pilot in Almaty.
Approximate render. Final look may differ.
01 · Why
For people who can't see. And people who can't look.
OKO is designed with blind users because the need is sharpest there. But the product is for anyone who needs to understand nearby space without pulling out a phone.
02
Warn about obstacles
A distance sensor on the glasses themselves gives a haptic alert about nearby obstacles - before and independently of the AI.
03
Read text
Signs, price tags, documents - text recognition runs on your phone, no cloud needed.
02 · System
Three layers. Each useful on its own.
The glasses are a thin sensor, the phone is the local brain, the cloud is an optional extra. The system degrades gracefully: lose internet, keep everything that matters.
Glasses aloneNo internet needed
Deterministic obstacle warnings from the distance sensor. A separate safety loop that depends on neither the phone nor the AI.
Glasses + phoneNo internet needed
Object recognition, text reading, voice answers - all local, on your phone.
CloudOptional
Richer scene descriptions when a connection exists. The cloud never controls the critical proximity alert.
Warns before the agent
The obstacle alert is a separate deterministic loop on the glasses. The AI agent can hang; the warning still fires.
One sentence, not a paragraph
A voice interface that respects attention: object, distance, direction - then silence.
03 · Limits
Assistance, not a promise of absolute safety.
We are honest about what OKO can and cannot do. Trust matters more than big claims.
Not a cane replacement
OKO is an additional orientation layer. It does not replace a cane, a guide dog, mobility training or professional guidance.
Camera on demand only
The glasses do not record continuously. Whenever the camera is active, a physical LED is visible to people around you.
Your data is not for sale
No advertising, no sale of sensory data. Core features run locally and stay free.
04 · Status
Project status
We build in the open. Here is the honest picture: what's done, what's in progress, what's next.
Done
- System architecture: glasses as sensor, phone as brain, cloud optional
- Component matrix: ~$34 and ~60 g per prototype
- 3D frame v0.3: three enclosure variants printed
In progress
- The phone brain: object recognition, speech, voice answers
- Glasses firmware: image streaming and the alert loop
Next
- End-to-end demo: question → answer in 2-3 seconds
- First pilot cohort in Almaty
05 · Join
Ways to join
There is nothing for sale yet - and we take no pre-payments. But there are four ways to be part of the project right now.
01
Waitlist
Want this device when it ships? Leave a contact - you'll hear first and get priority.
Get in line02
Almaty pilot
We're looking for blind participants, their families and accessibility specialists. Testing is free.
Apply03
Partnership
NGOs, rehab centers, universities, companies - let's run a measurable pilot together.
Write to us04
Funding
Grants, accelerators, investors - request the funding brief: team, plan, use of funds.
Request the brief06 · Waitlist
Waitlist
No pre-payment, no obligations. Just a way to say "I want this" - and hear about the launch first.