# CommonGround

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Daeyoung Kim](https://cerebralvalley.ai/u/kz3r0n3t)
- **GitHub:** https://github.com/kdy1029/common-ground
- **Demo video:** https://youtu.be/D5DiKoeiRho
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1554

CommonGround — An AI Interpreter That Understands What You're Pointing At

Traditional translation tools translate words, but they often miss the physical context behind them. When someone says "give me this one" or "plug that in here," the listener may still have no idea which object they mean.

CommonGround is a real-time, multimodal AI interpreter that combines speech, vision, and conversation context to help people communicate across language barriers.

CommonGround uses three Meta AI models. Muse Voice transcribes speech in real time. Muse Spark understands spoken references using camera frames and conversation history. SAM 3.1 identifies and highlights referenced objects in the video.

CommonGround maintains context across conversation turns, allowing it to understand references like "that one" or "the one from earlier." When a reference is ambiguous, it asks for clarification instead of guessing.

The working prototype supports live Korean-English video calls, with real-time translations, visual object highlighting, and graceful fallback when AI models fail.

Our goal is to make communication more natural and reliable in real-world situations, from remote assistance and education to hands-free collaboration through future wearable devices.

CommonGround: Translate when confident. Clarify when ambiguous.

## More from Meta Global AI Developer Hackathon

- [SaiGon Tour](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1551)
- [Xorshield](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1552)
- [ECO](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1553)
- [TYSA (Teach Yourself Anything)](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1555)
- [Amit S](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1556)
- [Ctrl Alt Muse](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1557)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1554. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt