ComfyUI Consistent Character Video Tutorial

cgpixel23 · reddit · 2026-07-12

A ComfyUI tutorial on generating character-consistent videos using the best Face ID LoRA. The author introduces a **Consistent Face-to-Video** workflow for low-VRAM users, focusing on a new face consistency LoRA to keep character identity stable across the video. The process requires only one reference image and a specific `ref_t2v` trigger word. It uses **GGUF models** to enable good face-to-video results on low-end devices. The post includes links to the article and video tutorial.

Original post →

More from Multimodal

Multimodal channel →