PROJECT 01 · 2024
Comubridge
A communication bridge for people who are hearing-, visually-, or speech-impaired.
Why it matters
Comubridge exists because most communication tools quietly assume everyone can hear, see, and speak the same way. This one doesn't. It's a full-stack app built so someone who signs, someone who reads text-to-speech, and someone typing can all end up in the same conversation without one of them doing extra work to be understood.
01
Why it matters
Accessibility isn't a feature request in Comubridge, it's the brief. The starting question was: what does a conversation look like when the three people in it are each missing a different channel , hearing, sight, or speech? Every screen had to hold up under at least one of those constraints.
02
How it's built
The frontend is React, talking to a Node.js service layer, deployed on AWS EC2. I focused early decisions on semantic structure and keyboard flow rather than visual polish, since that's what actually determines whether an assistive-tech user can complete a task.
03
What I'd extend next
Real-time captioning latency and offline fallback for low-connectivity users are the two gaps I'd close first if I kept building this out.
Outcome
Deployed and running on AWS EC2, with the interface built and tested around screen-reader flows and reduced-motion behavior from the start, not bolted on afterward.