Comic_Type

Comic_Type is the payoff node. The whole ComfyUI_StoryDiffusion pack exists to generate a batch of images where the same character appears across a sequence of scenes - and this is the node that takes that batch and lays it out like an actual comic page, captions and all. It's pure post-processing, no models involved, and it's the friendliest node in the pack.

You feed it the frames (the LATENT from StoryDiffusion_KSampler, decoded to an IMAGE), plus the scene captions, and it returns one assembled comic sheet. That's the whole job.

Inputs

• image - the batch of story frames. One frame per panel.
• scene_prompts - one caption per frame, one per line. Extra newline-free trick: if you give it a single line, it splits on the split_lines separator instead (default "；", a Chinese semicolon).
• fonts_list - the dropdown of fonts bundled with the pack. There's exactly one, Inkfree.ttf, a friendly handwritten style that fits comic captions.
• text_size (1–100, default 40) - caption size. 40 is a good starting point for 768px frames.
• comic_type - the layout. Four_Pannel arranges frames in classic 2×2 grids (padding with blank panels to a multiple of four). Classic_Comic_Style is a looser newspaper-strip layout that distributes frames across rows, some full-width, some stacked.
• split_lines - the separator for single-line captions, default "；".

Output

image - one IMAGE, the finished comic page. Straight into Save Image.

A genuinely nice detail: the node cleans the prompts itself before captioning. It strips the [Taylor]-style role markers, [NC] prefixes, parentheses and # comments from each line - the same markup StoryDiffusion_CLIPTextEncode used to build the scene. So you can paste the same scene_text you used for generation straight into scene_prompts and get clean captions without extra cleanup. It uses each line in order, and Four_Pannel requires the caption count to match the frame count, so keep your scene lines in sync with your batch.

Install

No extra dependencies beyond the pack itself:

cd ComfyUI/custom_nodes
git clone https://github.com/smthemex/ComfyUI_StoryDiffusion.git
cd ComfyUI_StoryDiffusion
pip install -r requirements.txt

The font ships inside the repo (fonts/Inkfree.ttf), so nothing to download. If you want more fonts you'd be editing the repo's fonts folder - the dropdown only lists what's there.

Honest take: it's a simple node, and that's fine. It won't win layout awards - Classic_Comic_Style can put a caption under an odd frame arrangement and there's no per-panel caption styling - but it's a one-click way to go from "five consistent images" to "a page that reads like a story", which is exactly what a beginner who googled this pack wants. If you outgrow it, export the frames and do the lettering in an image editor; until then, this saves you the copy-paste.