
OmniShow is an AI-powered video generation platform that creates studio-quality human‑object interaction (HOI) videos from a single product photo plus optional text, audio, and pose inputs.
Built on peer‑reviewed research (ByteDance, CUHK, Monash, HKU) and open‑sourced on GitHub, OmniShow is validated on the HOIVG‑Bench benchmark and outperforms baseline models in visual fidelity, motion naturalness, identity consistency, and condition alignment.