AtMoS
Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

OmNi-Ledger is a deterministic receiving and recall-control demo for pharmaceutical distributors. For this hackathon, I added a Muse Spark 1.3 adapter through the Meta Model API. It reads synthetic supplier-invoice text or an image and proposes structured fields; existing deterministic schema, catalogue, expiry and quantity checks decide whether the in-memory demo accepts or blocks the receipt. A valid synthetic sample is accepted and an expired sample is blocked as EXPIRED_GOODS. The model cannot authorize stock. This submission demonstrates the integration, not production readiness or measured OCR accuracy.