This is a workflow for the new H3 IMAGE model, not the h3 Video Gen model! See model download links below.
Text2Image / Image2Image / Ref2Image (9 refs max)
Pose Studio built in (toggle on and off to use it without ever switching workflows)
Flexible Image Comparison w/ multi-source toggles
Easy prompting template for highly detailed finetuned outputs
Grab images from videos using the Ref Loader (load video -> select frame -> "Use Frame" to have a snapshot be loaded in as an Image ref)
It comes with an organized layout, a wide variety of features and useful nodes, default templates to help get you started, and many other features.
Model Links
diffusion_models
text_encoders
https://huggingface.co/Comfy-Org/Qwen3.5/resolve/main/text_encoders/qwen3.5_9b_w4a8_gptq.safetensors (6.84 GB)
vae
Consider supporting me doing this full time & join my Skool community: https://www.skool.com/ai-filmmaking-w-foxfuressence-7902/about
Where I advocate for the use of 100% free, local, open source models and tools for creating character/story driven short films.
Join my discord! (AI Filmmaking discussions, help/troubleshooting, creation showcases, and more):
Description
Updated to v1.1 to add blurring to Pose Studio's output, allowing more natural poses rather than the rigid ones people were getting. I also improved the default prompt templates.