chat_bubble
Real-Time Scene Recognition
A short spoken summary of the space ahead — people, doors, seating, stairs, traffic state and free paths — refreshed on demand or continuously.
AI Vision Wearables See the world
See the world
through sound.
Sprhava AI-powered spectacles turn live camera input into clear spoken guidance — reading, recognising and warning in real time, hands free.
42 g
Frame weight
0.4 s
Scene response
9 h
Mixed-use battery
22
Voice languages
verified_user CE · FCC · RoHS documented
memory On-device neural processing
record_voice_over Open-ear bone conduction audio
public Distributor support in 30+ markets
What Sprhava Is
Unobtrusive access to the visual world
Sprhava is a wearable AI vision aid built for people living with blindness, low vision, macular degeneration, glaucoma or diabetic retinopathy. The frame looks like everyday eyewear — the intelligence sits inside the temples.
Instead of a bulky headset or a phone held at arm's length, wearers simply look and listen. A dual-camera module reads the scene, compact vision models interpret it, and a natural voice explains what matters — in under half a second.
- check_circleHands-free by design: tap, voice command or automatic mode — no screens, no pinch-zoom.
- check_circlePrivacy first: frames process core tasks locally; imagery is never stored by default.
- check_circleClinically aware: co-developed with low-vision therapists and O&M instructors.
descriptionReading mode"Boarding pass — Gate 14, seat 22A."
faceFace recognition"Maria is in front of you, two metres."
sensorsObstacle alert"Low branch ahead — step right."
shopping_basketProduct ID"Whole-grain pasta, 500 g, expires 2027."
Live audio output — illustrative
Core Capabilities
Six assistive functions, one pair of glasses
Each mode was prioritised from field interviews with more than 400 blind and low-vision users across daily-living, workplace and mobility tasks.
menu_book
Text-to-Speech Reading
Instant OCR for letters, menus, labels, packaging, screens and handwriting, with automatic language detection and adjustable reading pace.
travel_explore
Obstacle & Drop Detection
Depth-aware alerts for kerbs, poles, glass doors, head-height hazards and step-downs, announced by direction and distance as an O&M supplement.
category
Object & Colour ID
Find the right can, sort medication packs, match clothing colours or locate a dropped key — trained on 12,000+ household and retail items.
groups
Face & Expression Cues
Optional recognition of names the wearer has saved, plus a discreet cue when someone is facing or speaking to them.
forum
Ask-Anything Assistant
Conversational follow-ups — "Which queue is shorter?", "What is the expiry date?" — answered from the current field of view.
How It Works
From photons to plain language in four stages
An engineering pipeline tuned for latency, battery life and speech clarity — the three variables that decide whether an assistive wearable is actually used every day.
STAGE 01
Capture
Dual 12 MP wide-angle sensors with HDR and low-light gain follow the wearer's natural head direction — no aiming required.
STAGE 02
Interpret
An embedded NPU runs quantised detection, depth and OCR models locally; complex scene reasoning escalates to an encrypted cloud model.
STAGE 03
Prioritise
A relevance filter suppresses noise and speaks only what changes the wearer's next decision — safety first, then task, then context.
STAGE 04
Speak
Bone-conduction transducers deliver clear speech while ears stay open to traffic and conversation, with haptic cues for urgent alerts.
| Frame weight | 42 g, titanium-reinforced temples |
|---|---|
| Cameras | 2 × 12 MP wide-angle, HDR, low-light mode |
| Processing | Edge NPU + optional secure cloud inference |
| Audio | Open-ear bone conduction, dual beamforming mics |
| Response time | ≈0.4 s scene summary, ≈0.6 s full-page OCR |
| Battery | Up to 9 h mixed use; USB-C fast charge |
| Connectivity | Wi-Fi 6, Bluetooth 5.3, iOS & Android companion |
| Languages | 22 output voices, 40+ OCR scripts |
| Lenses | Prescription-ready, contrast filter options |
| Compliance | CE, FCC, RoHS, REACH documentation on file |
Engineering Detail
Built to be worn for nine hours, not nine minutes
Assistive wearables are abandoned for three reasons: weight, heat and chatter. Sprhava attacks all three. Thermal mass is distributed along both temples, keeping surface temperature under 39 °C in continuous scene mode, while a relevance engine cuts spoken output by roughly 60% compared with raw model narration.
Firmware updates arrive over the air with a staged rollback path, so clinics and distributors can validate a build before it reaches every unit in the field.
- boltOffline safety layer keeps obstacle alerts and short-text reading live without a network.
- lockLocal-only face library; encrypted at rest, wearer-deletable at any time.
- accessibility_newVoice, tap and single-button control paths for users with limited dexterity.
Where It Matters
Daily independence, measured in small wins
What our clinical partners report most often is not a dramatic transformation — it is the return of ordinary tasks people had quietly given up.
home
At Home
Sorting mail, reading dosage instructions, checking appliance displays and confirming expiry dates without waiting for help.
directions_walk
On the Move
Finding entrances, bus numbers, platform signs and empty seats; early warning for head-height and glass hazards.
work
At Work & Study
Reading printed handouts, whiteboards and badge names — a practical accommodation for employers and universities.
local_hospital
In Rehabilitation
Used inside low-vision therapy programmes as a graded training tool alongside magnifiers and O&M instruction.
For Buyers & Partners
Supplying accessibility programmes worldwide
Sprhava works with assistive-technology distributors, low-vision clinics, rehabilitation centres, disability-service tenders and OEM partners.
inventory_2
Distribution & Tenders
Volume pricing tiers, tender documentation, product data sheets, HS codes and lead-time commitments for public procurement cycles.
tune
OEM & ODM
Custom frame colourways, branded packaging, localised voice sets and white-label companion apps from moderate MOQs.
school
Clinical Enablement
Demo kits, clinician onboarding sessions, fitting protocols and printable user guides in large print and braille-ready files.
support_agent
After-Sales Service
Regional RMA handling, spare-part kits, 24-month limited warranty and firmware maintenance roadmap.
fact_check
Compliance Pack
CE, FCC, RoHS and REACH files, battery UN38.3 test reports, MSDS and declaration of conformity supplied per shipment.
language
Localisation
Voice output in 22 languages, OCR across 40+ scripts, and region-specific signage and currency models.
Field Experience
Tested by the people who rely on it
Feedback gathered during structured trials with rehabilitation partners in Europe, Southeast Asia and North America.
"The difference for our clients is latency. When a description arrives fast enough to act on, the device becomes part of mobility practice rather than a demonstration piece."
LM
Lena M.Orientation & Mobility Instructor
"I read my own post again. That sounds small. It isn't. I stopped asking my daughter to open my letters for me."
RA
Rohan A.Trial participant, retinitis pigmentosa
"Open-ear audio was the deciding factor for our clinic. Clients keep full environmental hearing, which our therapists insist on."
DC
Dr. ChenLow-Vision Rehabilitation Clinic
FAQ
Questions buyers and users ask first
Straight answers on capability limits, connectivity, safety and partner support.
The spectacles capture the wearer's field of view, analyse it with on-device and cloud computer vision models, then describe the result through open-ear audio. Wearers can read printed text and screens, identify products, recognise saved faces, locate doors or seats, and receive spoken alerts about obstacles ahead — entirely hands-free.
Yes. Obstacle detection, short-text OCR and colour or light detection run on the embedded neural processor with no network at all. Long-form scene description and conversational Q&A use a secure cloud model when Wi-Fi or a paired phone connection is available.
No, and we are explicit about that. Sprhava is a complementary electronic orientation and mobility aid. It adds context and early warnings, but primary mobility tools and training from a certified O&M specialist remain essential. We ship this guidance with every unit.
Yes. The frame accepts standard single-vision prescription lenses and contrast-enhancing filters, so low-vision wearers keep any usable residual sight while the AI layer fills the gaps. Fitting instructions are provided for optical partners.
Camera frames are processed and discarded by default — nothing is recorded unless the wearer explicitly saves an image. The face library stays encrypted on the device and can be wiped at any time. Cloud requests are transmitted over TLS and are not used to train models without documented consent.
Partners receive multilingual firmware, clinician training material, demo units, spare-part kits, full CE/FCC/RoHS documentation and OEM or ODM branding options, plus a named regional contact for technical and after-sales issues.
Resource Library
Explore assistive vision topics in depth
Technical explainers, comparison guides and buyer resources across the visual assistive technology category.
Bring AI vision assistance to your market
Speak with our product team about clinical evaluation units, distribution territories or OEM programmes for Sprhava AI-powered spectacles.
Contact Sprhava