OmniChar .char format adds consistent voice cloning from a 10-second sample in ComfyUI
ashishsanu · reddit · 2026-10-06
Developer ashishsanu extended the OmniChar .char portable character format with voice consistency: a 10-30 second mp3/wav sample is enough to keep a character's face, clothes, body, and now cloned voice consistent across multiple video generations, with lip sync handled natively by MiniMax H3.
Resources and limitations:
- Updated ComfyUI node (nightly) ships with open workflows on GitHub: encode a .char with voice, generate with MiniMax H3, plus full face/body/cloth workflows. The Omnichar repo is GPLv3 and supports krea2 and character finetuning.
- Prompt examples and a community character library are at omnichar.org/characters.
- Caveats: works best in English (non-English may slur), and avoid multiple voices in the sample.
Related event: OmniChar Adds Voice Consistency to .char Character Format(2 posts)→
More from Multimodal
- One-word Midjourney v8.2 prompt: 'Diaphanous' renders sheer translucency — tisch_eins · 2026-10-06
- Reddit User Shares AI-Reimagined 'Princess and Dragon' Short Video — Throwaway350750 · 2026-10-06
- Veda's sparse-attention ComfyUI node renders 5s video 2.9x faster on a 12GB RTX 5070 — lmoroney · 2026-10-06
- invideo launches MCP with 250+ SOTA models for full-stack AI video production — aziz4ai · 2026-10-06
- AI-Created Series 'Time Travel Tiffany' Drops Episode 2 — CavalierArcher · 2026-10-06
- Higgsfield claims every viral AI influencer is made on its platform, drawing scam-tool criticism — kyliebytes · 2026-10-06