Self-directed
Echo
accessibility
voice
Project overview
Audio-first navigation for blind web users
A background tool that interprets visual interfaces with computer vision and turns digital spaces into spoken, spatial sound cues.
Date
May 2026
Field
Infrastructure
Stack
Python, mss, PIL, Gemini Vision, Backboard
screen understanding
voice feedback
sound-led navigation
Context
Built for Husky Hacks 2026.
Most digital experiences still assume vision as the primary interface.
The hard part
Echo explores how web browsing, games, and visual software can become navigable through sound.
What it took
- Captured active screen state and converted it into navigable descriptions.
- Combined computer vision with spoken guidance for non-visual interaction.
- Explored game and web use cases beyond conventional screen-reader patterns.