EyeSense
Accessibility-focused Android visual assistant
- Status
- Mobile prototype
- Engagement
- Innovation showcase project
- Role
- Full-stack & AI engineer
The problem
Blind and low-vision users need practical visual assistance that respects privacy and works when connectivity is limited.
What was built
A Flutter Android prototype with offline OCR, speech and vision assistance, a Cloudflare Worker backend, and a TalkBack-first design that asks for consent before any image leaves the phone.
Architecture and engineering decisions
Flutter and Dart on Android with on-device OCR and speech, plus a Cloudflare Worker for optional cloud vision calls. Consent is requested before any image is sent off-device.
Audience
Blind and low-vision Android users.
Technical highlights
- Offline OCR and speech assistance
- TalkBack-first Android design
- Consent before images leave the device
- Cloudflare Worker for optional vision calls
My responsibilities
- Mobile implementation
- Accessibility design
- On-device OCR integration
- Consent flow
Why it matters for a client
Shows accessibility-first mobile engineering and privacy-aware handling of camera data.
Constraints and trade-offs
A prototype, not a validated navigation aid. Performance and accuracy depend on device capabilities and the quality of captured images.
Privacy and accessibility
Consent before any image leaves the phone; TalkBack-first interaction. No claim of formal accessibility certification.
Verified evidence
Public source at github.com/RealDevRay/eyesense. Described on the owner portfolio as an accessibility prototype with offline OCR and consent-gated image handling.
Limitations
Prototype scope; not a validated navigation aid and not a substitute for existing mobility tools. No App Store availability claimed.