- Type
- Generative AI
- Year
- 2026
- Tools
- ComfyUI, Flux.2, LTX 2.3
The project
The goal: building a facial animation pipeline synced to an audio track, with natural, expressive lip sync that keeps the character's high visual fidelity.
Sync workflow
01Character design
Creating the face and environment with Flux.2, for a photorealistic base.
Creating the face and environment with Flux.2, for a photorealistic base.
02Video animation
LTX 2.3 generates the micro-movements that make the sequence feel natural.
LTX 2.3 generates the micro-movements that make the sequence feel natural.
03Audio-driven animation
A lip sync workflow in ComfyUI, based on LTX 2.3: the soundtrack drives the mouth movements and facial expressions.
A lip sync workflow in ComfyUI, based on LTX 2.3: the soundtrack drives the mouth movements and facial expressions.
Tools and skills
- ComfyUI
- Flux.2
- LTX 2.3
- Audio input
Key points
01
Realistic phonemes
Lip movements are precisely matched to the provided audio.
02
Identity preservation
The character's face stays consistent throughout the animation.
