CommunicationTestingUpdated Jul 26, 2026
ASL Speaker
Conversations between signers and non-signers usually depend on a third person being present to interpret.
Why this matters
Direct conversation shouldn't require an interpreter to always be available. A real-time bridge between ASL and spoken language creates more spontaneous, everyday interactions.
- Current idea
- ASL Speaker translates American Sign Language gestures into spoken words in real time, using computer vision and machine learning to recognize signs and vocalize their meaning.
- Accessibility notes
- Visual-first feedback with clear sign detection, large output text, and adjustable speech rate.
- What was learned
- Real-time translation is only useful if it is fast enough to feel like a conversation, not a presentation.
Development updates
Progress in public
Real-time sign recognition and spoken output prototype released.
Computer vision pipeline and initial ASL vocabulary model tested.
Feedback
Have thoughts on this project?
Head to the Ideas board to share a related barrier or idea. Attribution is optional.
Open the Ideas board