ComfyUI Node Runs on cloud

Image Padding

Add breathing room around an image (and get it back cleanly)

By WASasquatch·Created 3 years ago·Updated about a year ago· 1,812
Image Padding

      Image Padding adds a border of extra space around your picture. That sounds trivial until you realize what it's for: giving a diffusion model room to paint into. The single most common use is outpainting - you pad an image outward, hand the padded canvas plus a mask of the new area to an inpainting pass, and the model extends the scene into the space you created. No padding, no room to grow the frame.

      WAS Node Suite is, as the ecosystem notes put it, "a large collection of image processing utilities" that fills in the post-processing gaps core ComfyUI leaves open, and padding is one of those quietly essential gaps. It also shows up outside outpainting: adding a margin before a rotation so the corners don't clip, giving a subject some breathing space before a composite, or matting an image to a target canvas size.

      How it works

      It places your original image onto a larger canvas and fills the new border region - that's the padding. In an outpainting flow the padded border is exactly the area you then mark as "generate here," which is why padding and masking go together. Nodes in this family typically also hand back the coordinate data describing where the original sat inside the padded canvas, so you can crop back to the exact original region later if you need to. That round-trip - pad out, do the work, crop back - is the pattern to hold in your head.

      The inputs and outputs that matter

      • The image in - what you're padding.
      • A padding amount - how much space to add. In an outpainting setup this is how far you're extending the frame.

      Heads up on honesty: this node isn't described in the WAS README's node list, and I don't have a schema exposing its exact fields, so I'm describing the function rather than naming specific widgets, defaults, or every output port. What it does - surround an image with fillable space, usable as an outpainting canvas - is solid; the precise labels you'll read off the node itself once it's on your canvas.

      How to install it

      It's part of the WAS Node Suite:

      • ComfyUI Manager: search was-node-suite-comfyui, install, restart.
      • Manually: cd ComfyUI/custom_nodes && git clone https://github.com/WASasquatch/was-node-suite-comfyui/, then pip install -r requirements.txt against your ComfyUI Python, and restart.

      Padding itself is a plain pixel operation with no model download. The heavy work in an outpainting flow lives in your inpainting model, not here.

      Common issues

      The gotcha isn't the padding, it's the seam. When you generate into the padded region, the boundary between the original pixels and the freshly-painted ones can show as a visible line or a tone shift. The fix is the standard inpainting hygiene the KB harps on: soften the mask edge (feather it), let the generation overlap slightly into the original, and match your denoise so the new content blends rather than butts up against the old. Padding gives you the canvas; blending is what sells it.

      Watch your final dimensions, too. Padding makes the image bigger, and if a downstream node or your sampler expects a specific size, you'll want to either crop back to the original region afterward or resize deliberately - don't let a padded canvas silently blow up your output resolution.

      And the pack-wide caveat: WAS Node Suite has been retired since December 2023 - still used everywhere, just unmaintained - so if a ComfyUI update makes every WAS node vanish with an "Import Failed", that's suite-wide dependency drift, not this node. Re-run requirements.txt against the right Python environment (or the bundled install.bat) and restart.

      Categoryx

      Inputs (0)

      No inputs

      Outputs (0)

      No outputs