Open-Source ComfyUI Workflow Enables Ultra-Long Lip-Sync Talking Head Videos

A Reddit user has open-sourced a ComfyUI workflow based on MiniMax H3 that generates ultra-long lip-synced talking head videos from audio, reference images, and keyframes. It uses segmented generation and stitching to prevent quality degradation, with the code available on GitHub.

2026-10-27 ~ 2026-10-27 · 2 related posts

• Redditor open-sources a simple workflow for infinite lipsynced talking head videos with MiniMax H3 — martinerous · 2026-10-27
• Open ComfyUI Workflow: Long Lipsync With MiniMax, Segmented to Avoid Degradation — martinerous · 2026-10-27