SignSpeak

Real-time sign language captioning for any video call.

Accra, Ghana 🇬🇭

What I’m building

SignSpeak watches a webcam and turns sign language into live captions the hearing people on the call can read — and turns their speech into text the signer can read back. It’s the captioning layer video calls forgot to build for deaf users.

Why I’m building it

I’m a deaf engineer. Every standup, I’m either waiting on a human interpreter’s schedule or reading laggy auto-captions of hearing people while my own signing goes unseen. Nothing about us without us — so I’m building it.

What the tokens unlock

Credits funded the compute to train the sign-recognition model on community-contributed video and the low-latency inference that keeps captions in sync with the conversation.

✓
Train the sign-recognition model on community video1.5M
✓
Low-latency live inference pipeline1M
✓
Ship the Zoom/Meet integration500K

Tools & models

Vision (pose)Sign→text modelCompute

kwamem/signspeak

The builder

Kwame Mensah

Accra, Ghana 🇬🇭

Deaf software engineer building the captioning layer the web forgot. Nothing about us without us.

Identity verifiedGitHub connectedOpen-source contributor

Updates

Fully funded — thank you
1d ago
We hit 106% of the goal. First public build lands next month. This is the first tool built for my calls, not around them.

Supporters

58 people have pledged so far.

Anonymous supporter· 2d ago
“Rooting for you. This should exist.”
500KOpenAI
Anonymous supporter· 4d ago
250KCompute
Rae ColemanChicago, USA· 6d ago
“Take my credits and go build it.”
1MClaude
Ken WatanabeTokyo, Japan· 8d ago
500KGemini
María TorresBarcelona, Spain· 10d ago
100KAI Credits

Discussion

Ask the builder a question or leave encouragement.

No comments yet. Be the first to say something.