signcall-hack-nation.jpg
SignCall – Hack-Nation Global AI Hackathon
2nd place in the ElevenLabs track at Hack-Nation 2026: an AI agent that turns sign-language video into real phone calls, so Deaf and non-speaking people can book appointments on their own.
Context
The 4th Hack-Nation Global AI Hackathon (February 2026) brought together 1,500+ builders across hubs at Stanford, Zurich, Munich and beyond, selected from 2,500+ applicants. Heidi competed at the Munich hub, hosted by the Center for Digital Technology and Management (CDTM), in the ElevenLabs track, with teammates Bethany Anne Wong, Hassan Sarwat and Aly Assem.
The problem
A phone call is still the default way to book a medical appointment or a restaurant table, and it shuts out Deaf and non-speaking people.
What we built: SignCall
An autonomous agent that restores independence to Deaf and non-speaking individuals:
- The user signs into a camera; the agent understands the intent from sign-language video.
- It then uses AI to place a real-time phone call on the user's behalf, for example to book a medical appointment or make a reservation, entirely on their own terms.
Tech
Sign-language video input · AI intent understanding · real-time voice calls powered by ElevenLabs
Why this shows I can…
Build accessibility technology with real-world impact, combine vision and voice AI into one agent, and deliver a demo that works under hackathon pressure.
Result
🥈 2nd place, ElevenLabs track — Hack-Nation Global AI Hackathon 2026. Out of 2,500+ applicants, the team was selected for the Venture Incubation Program, a dedicated track to evolve the prototype into a market-ready startup. ElevenLabs supported the team with API credits.
Thanks
To ElevenLabs for the challenge and the API credits, and to Kai Nestor Wiederhold, Linn Bieske and CDTM for hosting the Munich hub.