# gamAI

- **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:** [Renato Schuster](https://cerebralvalley.ai/u/renatogama534)
- **GitHub:** https://github.com/gamai-io/floor-copilot
- **Demo video:** https://github.com/gamai-io/floor-copilot/releases/download/v0.1.0/floor-copilot-demo.mp4
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1633

Floor Copilot is an evidence-gated invoice-intake copilot for small industrial distributors. A phone photo of a supplier invoice becomes structured line items; the app recomputes qty x unit price, net total, 17% VAT and total with Decimal math. If any number does not reconcile, the result is marked unverified with an exact problem list instead of a confident guess. Every result stores a SHA-256 of the image and an evidence record behind an unguessable token. Part inspection is wired to Meta SAM 3.1 and fails closed (0 masks, unverified) when no SAM model is configured. 7/7 unit tests run without a model or network; CI is green. Honest limits: photos only (no PDF), no field-level OCR confidence yet.

## More from Meta Global AI Developer Hackathon

- [StepProof Studio](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1630)
- [Sightra](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1631)
- [ReelDone](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1632)
- [Nyx](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1634)
- [Loss function](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1635)
- [Israel Lee Armstead](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1636)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/1633. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt