# Hydra AI

- **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:** [Georgio Saliba](https://cerebralvalley.ai/u/georgio_saliba), [Amine Ali Chaouche](https://cerebralvalley.ai/u/amineali)
- **GitHub:** https://github.com/georgiosaliba227/Hydra-Ai
- **Demo video:** https://youtube.com/shorts/umgbylpo-Hc
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4970

Hydra AI combines AI-powered beverage recognition, fluid intake history, and experimental personalized workout forecasting to help people understand their drinking habits and exercise-related fluid losses.

Built as a native iPhone app, Hydra lets users photograph a drink, review its suggested identity and supported container size, and confirm how much they consumed. Our locally adapted Meta Muse Glimmer model powers beverage recognition through a Mac backend. Users can edit suggestions, and personal correction memory preserves their preferred drink details. Uncertain sizes remain unknown until the user supplies them.

Beyond logging, Hydra uses comparable personal workout measurements to generate conditional fluid-loss estimates with transparent uncertainty. Daily totals, editable history, user-chosen goals, and sourced educational answers bring these capabilities into one experience. Numerical results come from deterministic calculations.

We have also implemented a Meta Muse Wonder integration for explaining calculated trends and relevant context. That integration has been tested with mock providers; live validation remains pending.

Our vision is a personalized wellness companion that connects intelligent drink logging, personal workout calibration, Apple Health, and weather context. Automatic workout import and weather integration remain outside the current demonstrated workflow. Forecasting remains experimental, with estimates clearly distinguished from measured outcomes and daily drinking requirements.

## More from Meta Global AI Developer Hackathon

- [Jarvis](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4967)
- [solehero](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4968)
- [MiloBroker](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4969)
- [Timosgroup](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4971)
- [Cobweb](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4972)
- [Mohammad Yehya](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4973)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4970. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt