# Verity — Nnaemeka Eziamaka

- **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:** [Nnaemeka Eziamaka](https://cerebralvalley.ai/u/chidelu)
- **GitHub:** https://github.com/EziamakaNV/verity
- **Demo video:** https://youtu.be/52kxwbvfEww
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4198

Verity — Know the story behind your product.

Counterfeit consumer products in Nigeria can arrive in convincing packaging, leaving shoppers unable to assess a seller’s explanation. Verity helps a buyer inspect the recorded journey of an individual product and investigate inconsistencies before paying. A copied QR code alone is never treated as proof of authenticity.

Built during the hackathon: a React/TypeScript consumer interface, QR camera and image-upload decoding, a partner workflow for custody, sales, linked returns and restocks, a Cloudflare Workers/D1 backend, and a live Muse Wonder investigation with inspectable evidence and an actual retrieval trace.

The demo contrasts two cases: a previously sold item offered again with a seller-supplied return document that refers to another serial and sale, versus a legitimate return and restock linked to the original sale. Exact identifier, chronology and lifecycle checks support the verdict; Muse retrieves relevant records, evaluates the seller’s explanation, and recommends evidence-backed next checks. The interface distinguishes the model’s statement from deterministic record checks.

Engineering includes atomic partner updates, idempotent retries, isolated demo sessions, evidence-snapshot validation, stale-result invalidation, bounded tool calls, citation validation and explicit live-provider failure states. The recorded live investigation made five retrieval calls across three Meta API requests in approximately 14 seconds.

All products, businesses and supply-chain records in the demo are fictional. The model investigation is live. Verity does not certify that a physical item is genuine or safe. A real deployment depends on manufacturer and retailer participation; the next step is a small pilot to test record quality and buyer usefulness.

Try the live demo: https://verity.eziamakanv.workers.dev
Select “Seller says it was returned”, then “Review evidence”. Compare with “A return that checks out”.

## More from Meta Global AI Developer Hackathon

- [Bonisa](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4195)
- [Kaypro](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4196)
- [floodwatch](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4197)
- [Ale + Jolly — Physical AI for Business](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4199)
- [Voni](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4200)
- [TransitionBio](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4201)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/4198. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt