VOCAVID

COMFYUI · WORKFLOW REFERENCE

The workflows.
No guesswork.

This overview is derived from the workflow JSON files currently bundled with VocaVid. Filenames and node types are therefore concrete; when loading a workflow, ComfyUI-Manager is still the most reliable way to identify missing custom nodes.

How to proceed: Download JSON → load it in ComfyUI → install missing nodes with ComfyUI-Manager → place the model files listed below in the model folders expected by the workflow. The JSON files remain the authoritative source because model and node dependencies can evolve.

AVATAR → IMAGE · QWEN

Qwen Image Edit

Creates or varies avatar and reference images with Qwen Image Edit.

  • qwen_image_edit_2511_bf16.safetensors
  • qwen_image_vae.safetensors
  • qwen_2.5_vl_7b_fp8_scaled.safetensors
  • Qwen-Image-Edit-2511-Lightning-4steps-V1.0-bf16.safetensors
  • Node types include TextEncodeQwenImageEditPlus, FluxKontextImageScale and ComfySwitchNode
Download avatartoimage.json

AVATAR → IMAGE · FLUX

Flux 2 Klein

An alternative avatar-to-image pipeline based on Flux 2 Klein.

  • flux-2-klein-base-9b-fp8.safetensors
  • full_encoder_small_decoder.safetensors
  • qwen_3_8b_fp8mixed.safetensors
  • Node types include ReferenceLatent, EmptyFlux2LatentImage and Flux2Scheduler
Download avatartoimage_flux.json

IMAGE + AUDIO → VIDEO

LTX 2.3

Combines image, audio and prompt into an animated clip.

  • ltx-2.3-22b-dev-fp8.safetensors
  • ltx-2.3-spatial-upscaler-x2-1.1.safetensors
  • gemma_3_12B_it_fp4_mixed.safetensors
  • ltx_2.3_22b_distilled_1.1_lora_dynamic_fro09_avg_rank_111_bf16.safetensors
  • Node types include CreateVideo and LTXAV nodes
Download imageaudiotovideo.json

IMAGE GENERATION

Z-Image Turbo

Fast image generation for storyboard and scene images.

  • z_image_turbo_bf16.safetensors
  • qwen_3_4b.safetensors
  • ae.safetensors
  • The JSON contains direct model hints and expected ComfyUI model folders.
Download image_z_image_turbo.json

PROMPT GENERATION

Gemma 4

Creates or structures scene and video prompts.

  • gemma4_e4b_it_fp8_scaled.safetensors
  • Node types include TextGenerate, GetVideoComponents and PreviewAny
Download promptgen.json
Set-up guide →FAQ with timings →VocaVid on GitHub →