ZF 简易尺寸

Sometimes you don't need an image resized - you just need two numbers. That's the entire job of ZFEasySizeSettings (display name "ZF 简易尺寸", "easy size"). Pick a preset dimension or type a custom one, and it outputs clean width and height integers you can wire straight into your model's native empty-latent node. No pixels get touched, no latent gets created, nothing else happens. That's the whole node, and that's why it's useful.

It's part of ZF-ComfyUI-Helper, the dependency-free utility pack, under ZF Helper/Size.

Why a node that only outputs numbers exists

Because in ComfyUI, resolution is something you repeat constantly, and repetition is where bugs come from. You pick "1024x576" in a preset, type it into an empty latent, type it again into a resize node, and the third time you change it you forget one of them - silent mismatch, weird results, confusion. ZFEasySizeSettings gives you one authoritative source for the numbers and fans them out: connect width and height to an EmptyLatentImage, to a resolution-conditioning node, to anything that takes integers, and you change it in exactly one place.

The preset list is the same model-neutral one as the pack's image-resizing node - deduplicated and regrouped from SD1.5/SDXL/FLUX/WAN/QWEN resolution tables into landscape, portrait, and square submenus. If your model family is FLUX or WAN, you'll recognize the sizes, but nothing here is locked to a model. Every preset is a multiple of 8, which most modern latent and video models appreciate.

The inputs and outputs

Only four inputs, and you'll normally set two:

• size_mode - preset (pick from the dropdown) or custom (use width/height).
• preset - the dropdown, when in preset mode.
• width / height - used in custom mode. Ranged 64–8192, stepped by 8, so you can't accidentally produce a weird non-multiple-of-8 size.

Three outputs: width, height, and summary - a string like 1024x576 | 16:9 | mode=preset you can route to a text-display node if you want the resolution visible on the canvas.

Installing it

The whole pack is dependency-free, so it's the standard quick install:

cd ComfyUI/custom_nodes
git clone https://github.com/Z-yaofang/ZF-ComfyUI-Helper

Restart ComfyUI, or install via ComfyUI Manager by searching ZF-ComfyUI-Helper. No models, no requirements.txt.

The one thing to keep straight

This node is numbers only. It does not create a latent, so the common wiring is: ZFEasySizeSettings → EmptyLatentImage → sampler. The pack's sibling ZFEasySizeImage is the one that actually resizes an image (and its mask). If you connect this node to an empty latent and the image you're conditioning on is a different size, that's on you - this node never looked at any pixels. That's not a bug, it's the design: one source of truth for the numbers, and you decide where they land. If you're tired of typing resolutions into three separate places per workflow, this is the node that kills that habit for good.