FREE vs FULL — Start free: https://whop.com/videolab-073d/free-starter-fasth3-omni-lip-sync
Use code FAST30 for 30% off ($27.30).

Every talking-head tool either mutes your character or charges you per second. FastH3 Omni does it in one pass: one reference photo, your own voice line, and you get a 5-second video of your character speaking your exact words — with sound built in.

WHAT’S INSIDE
• The ready lip-sync preset (ComfyUI workflow): load portrait + voice line, run, done
• The 3 official FastH3 templates (reference-to-video, image-to-video, text-to-video)
• The exact settings that matter: 8 steps, Euler/simple, sigma shift 12/3, VSA 10%, 5 s = 124 frames, 0.6 MP
• A prompt library with 10 copy-paste formulas: talking portrait, product presenter, storyteller, narrator, reaction shot and more
• Quick-start guide (EN) + full workflow doc (AR)

A GPU is required (Comfy Cloud or a GPU PC — it will not run on a CPU-only laptop). Model files are downloaded free and are NOT in the ZIP (about 40 GB). Licence note: MiniMax H3 open weights are under the MiniMax H3 Community License Agreement — self-hosted use is NOT licensed in the US, EU, UK or South Korea without applying to MiniMax first. Check the licence for your country.

For: creators who want talking-head videos, product presenters and UGC-style clips without a camera or per-second subscription. NOT for: anyone without a GPU, or anyone wanting a long video in one click — this builds 5-second clips you join together.