Kanu · wearable AI guide · by Code-X
It works offline and always gets you home. Kanu sits on your shoulder, sees all around you and tells you where to turn with a buzz on each shoulder. It is built first for people who are blind or have low vision, in Arabic and English, with all its AI on the device.
Drag to orbit · Click a part to isolate it
The 3D model needs WebGL. Open this page in a current version of Chrome, Edge, Safari or Firefox. The sections below describe both editions.
Concept model. Kanu Companion follows the V1 prototype. Kanu Guide shows the design for the next prototype.
The problem
Millions of people learn to move with a cane or a guide dog. What they still lack is something that warns about obstacles at head height, remembers a route, and says where to turn, without covering their ears or costing as much as a car.
Prevalence: GBD 2019 Blindness and Vision Impairment Collaborators, The Lancet Global Health, 2021, 2020 estimates. Prices: see the market table below.
Meet Kanu
Both editions share the same body, strap, battery and on-board AI. What sits on top changes, and so does the software.
For blind and low-vision users · Lead product
Headless and hands-free. A low 360° camera dome replaces the moving head, and two buzz pods, one on each shoulder, give directions you can feel.
For everyday life, travel and the outdoors
The edition V1 is built on: a head that turns to look, a face that reacts and a voice at your ear. Personas change what it listens for and says.
How Kanu Guide works
Say “save this place” or press the side button. Kanu records the walk from GPS, motion sensors and what its cameras see along the way.
The 360° dome and depth sensor watch head and chest height, the zone a cane cannot reach, and call out only what matters.
A buzz on the left shoulder means turn left. On the right, turn right. Your ears stay free for traffic.
Say “take me home”. Kanu retraces the saved route step by step, with or without a signal.
Six patterns, learned in minutes. Tap one to feel it on the model.
Where we are
The first prototype was designed in CAD, printed and assembled by the Code-X team. It rides on the shoulder with a cross-body strap, turns its head on servos and shows a face on its screen.
Photos and video are of the real V1, colour-corrected. In the studio shots the face on the screen is simulated.
The intelligence stack
Kanu’s software is modular. Each edition switches on the systems it needs, and new skills arrive over the air.
Recognises familiar people by face and voice, even in a busy room.
In developmentReads tone of voice and facial expression together to choose how to respond.
In developmentOffline translation of speech and signs, Arabic dialects first, drawing on Code-X’s Lebanese.ai dialect engine.
In developmentObstacle awareness all around, with priority on head and chest height.
In developmentStores paths from GPS, motion and visual landmarks, and leads you back along them.
In developmentOrchestrates every system and the hardware in real time, with no cloud round trip.
In developmentPrivacy
A device on your shoulder sees your whole day. Kanu is built so none of it leaves the device.
Vision, speech and route memory run on the device, with or without a signal.
Camera frames and audio are processed on the device and never uploaded.
The network brings in new models and skills. It never carries out what Kanu saw.
Why it is hard to copy
We are preparing patent claims around these areas. Each ties the body, the sensors and the software together, which makes it hard to bolt onto an existing product.
A direction language delivered through a pod on each shoulder, linked by a band, so turns are felt on the side you should go.
Routes stored on the device from GPS, motion sensing and visual landmarks, replayed step by step to lead the wearer back.
A 360° camera dome and forward depth sensor at shoulder height, covering the zone a cane misses.
The same shoulder saddle carries either a 360° dome or a pan-tilt head with a face, so one hardware line serves several markets.
On Companion, the head turns toward whoever is speaking, combining voice direction with the face in view.
The market gap
Today’s leading mobility aids are impressive, but they are priced in the thousands, weigh a kilogram or more, or leave Arabic speakers behind. Kanu is aimed at the gap between a smart cane and a premium headset.
| Device | Form | Price | Weight / battery |
|---|---|---|---|
| WeWALK Smart Cane 2 | Smart cane handle | Financing from $70/month | Cane |
| Glide (Glidance) | Wheeled guide you hold | $1,499 + $30/month | About 4 kg |
| NOA (biped.ai) | Harness over the shoulders | About £3,000 (UK) | Not published |
| OrCam MyEye 3 Pro | Camera on your glasses; reading and recognition | $4,250 | Clip-on |
| .lumen | Head-worn guide | €9,999 | About 1 kg · about 2 h |
| Kanu Guide | Shoulder wearable with buzz pods | Accessible price, set in pilots | Target under 300 g · all day |
Competitor figures are from each company’s site or published coverage in 2026 and may change: WeWALK, Glidance, NOA (user reports), OrCam, .lumen. Kanu figures are design targets.
Accessories
Roadmap
Wearable prototype with servo head, face display, speaker and strap.
360° dome, buzz pods and back band on the V1 body. Route memory on the device.
Trials with blind and low-vision users and partner organisations in Lebanon and the GCC.
Certification, manufacturing and launch with the Guide Starter and Plus bundles.
Platform vision
Team
Founder and CEO
Entrepreneur and business consultant with more than 20 years of experience. Founder and CEO of SourceX Holdings, a group of nine ventures across AI, software, marketing and commerce.
Engineering
The SourceX company for AI solutions, software development and cybersecurity. Code-X designs, prints and programs Kanu in-house, and builds Lebanese.ai, its Arabic dialect engine.
We are looking for investors, pilot partners among organisations for the blind, and accessibility advisors.
Get in touch
Book a private demo of V1, request the investor deck, or talk to us about a pilot.