Image Size Creator Plus ✨

Image Size Creator Plus is the same "give me the numbers" trick as the pack's plain Image Size Creator 📏, with one addition: a style selector that adds the Midjourney standard aspect ratios to the SDXL ones. So when you're matching a reference composition - a 16:9 crop, a 2:1 banner, a 5:6 feed image - there's a preset instead of a calculator.

It's part of the Painting Coder Utils pack, shipped in v0.3.5 as the enhanced sibling of the base creator, and it inherits the same behavior: it computes width and height from a ratio dropdown and returns them (plus a resolution string) without generating anything.

How it works

style flips the resolution catalog:

• SDXL - the nine training-optimal ratios: 1:1 (1024×1024), 9:7 (1152×896), 7:9, 3:2 (1216×832), 2:3, 7:4 (1344×768), 4:7, 12:5 (1536×640), 5:12.
• Midjourney - its standard set: 1:2 (768×1536), 9:16 (816×1456), 2:3 (896×1344), 3:4 (928×1232), 5:6 (992×1200), 1:1, 6:5 (1200×992), 4:3 (1232×928), 3:2 (1344×896), 16:9 (1456×816), 2:1 (1536×768).

Pick mode (Landscape / Portrait / Square) and the dropdown filters to matching orientations. scale_factor (0.1–10) multiplies the base size. Outputs are resolution, width, and height - pure numbers, ready to feed a latent creator, an image resize, or a conditioning area so every node in the graph agrees on the canvas.

Inputs and output

• mode - Landscape, Portrait, or Square.
• style - SDXL or Midjourney.
• resolution - the merged preset list (19 options after de-duplication).
• scale_factor - optional multiplier.
• Outputs: resolution, width, height.

Where you'll use it

When you want the SDXL-safe ratios and occasionally need a platform-style crop without switching nodes. If you only generate square or the core SDXL ratios, the plain creator covers you and Plus is just noise. Reach for Plus when composition matching matters - e.g. you're reproducing a Midjourney image's framing at a size your local model will actually respect.

Like its parent, it's metadata-only: nothing is generated. For an actual empty latent use Image Latent Creator Plus ✨, which builds on this one.

Install

ComfyUI Manager → Painting Coder Utils, or:

cd ComfyUI/custom_nodes
git clone https://github.com/jammyfu/ComfyUI_PaintingCoderUtils.git

Restart ComfyUI. No models or keys; dependencies (numpy, Pillow, torch, gradio) are already in any working ComfyUI. And the standing caveat for this pack: v0.3.0 renamed the node namespace, so older workflow JSON may not resolve - repair it with the bundled docs/fix/workflow_fixer.html.