Nodes/ComfyUI-Lora-Pipeline/Conditioning Area Preview
ComfyUI Node

Conditioning Area Preview

Finally, a way to see the areas you're conditioning

By andreszs·Created 5 months ago·Updated 12 days ago· 20
Conditioning Area Preview
  • areas
  • preview
width512
height512

Area conditioning in ComfyUI has always had a visibility problem. The built-in Set Area node gives you four number boxes and no picture, which is why "changing the coordinates does nothing" is a recurring community refrain - half the time people can't even tell if their region is where they think it is. Conditioning Area Preview is that missing picture. Feed it the area metadata from Conditioning Pipeline (Combine) and it renders a colored rectangle map of every region, so you can finally look at what you're conditioning instead of squinting at floats.

How it works

It's a pure diagnostic node - no sampling, no conditioning math. It takes the CONDITIONING_AREAS list (the areas_out output from Combine), draws each area as a color-coded rectangle on a canvas, and overlays a grid, crosshair axis, and a label per region showing its normalized coordinates (Area 1 · x=0.25 y=0.10 w=0.50 h=0.80). Small areas fall back to a short Area N badge so the label fits. The grid style intentionally matches the author's companion OpenPose Studio canvas, so the same visual language carries over when you mirror these areas into the pose editor. If the areas list is empty, you get a flat dark canvas - which is itself useful diagnostic information.

The inputs that matter

Only three, and none of them are tunable:

  • areas - wire areas_out from Conditioning Pipeline (Combine) in here. This is the metadata, not a preview of your image.
  • width / height - the render canvas size in pixels, default 512×512, range 64–4096. Note these are just the resolution of the preview image; they have zero effect on generation. Set them to something close to your target output so the proportions read correctly.

Output

One IMAGE (preview) - drop it into a standard PreviewImage or SaveImage node as a side branch. It's not on the sampling path, so it costs you nothing at generation time; it's the thing you look at while you're still placing areas.

Install

It ships as part of andreszs/ComfyUI-Lora-Pipeline - no separate install. ComfyUI Manager → search "LoRA Pipeline", or:

cd ComfyUI/custom_nodes
git clone https://github.com/andreszs/ComfyUI-Lora-Pipeline

Restart ComfyUI and it's under LoRA Pipeline → Conditioning. Zero extra dependencies: it draws with Pillow, which ComfyUI already requires, so there's nothing to pip install here. One quirk worth knowing: this node isn't actually listed in the pack README's node table (that documents Set Area, Combine, and the scheduled LoRA loader), so it's easy to assume it doesn't exist. It does - it's just the pack's hidden extra.

Common issues

  • Black preview. areas came in empty - usually means Combine got no valid pipeline entries (check your Set Area chain) or you wired the wrong thing in.
  • Areas in the wrong place. That's your coordinate math, and now you can see it. Remember Set Area's x/y are normalized 0–1 with the origin top-left; the preview shows the boxes exactly as the pipeline defines them, so if it looks off in the preview, fix it before you render, not after.

For the price of one wire, this node removes the blindest part of regional conditioning. If you're running the Lora-Pipeline area stack at all, it's worth the thirty seconds to add.

CategoryLoRA Pipeline/Conditioning

Inputs (3)

NameTypeDefaultDescription
areasCONDITIONING_AREAS
widthINT51264–4096
heightINT51264–4096

Outputs (1)

NameTypeDescription
previewIMAGE