ArchAi3D_Solid_Color_Image

Sometimes you need a flat, colored canvas that's exactly the right size for a model - a blank panorama to inpaint a room into, a background plate for compositing, a resolution test target for VRAM budgeting. ArchAi3D_Solid_Color_Image is a 295-line utility that does precisely that and nothing else. Its party trick: dimension presets that are all 2:1 panoramic aspect ratios, all divisible by 32, sized for Qwen-VL and the pack's panorama workflows, from 1024×512 (0.52 MP) up to 3072×1536 (4.72 MP).

The default preset is 2:1 - 768 (1536×768) - a size the author clearly picked as the sweet spot for interior-design editing with the pack's Qwen encoder nodes. If you're building a workflow around those, matching the preset to what the encoder was tuned for saves you an explicit resize.

How to use it

Pick a color two ways, via color_mode:

• rgb (default) - rgb_r/rgb_g/rgb_b sliders, 0–255 each, default mid-gray 128
• hex - hex_color string, accepts 808080 or #808080

Dimensions come from either source: connect reference_image (optional IMAGE) and the output matches its dimensions, or leave it disconnected and preset_dimensions takes over. That reference path is the sneaky-useful bit - it's a one-wire way to build a color canvas exactly matching another image's resolution.

debug_mode (default False) turns on console logging of the dimensions and color. Output is a single IMAGE (solid_image).

Installing it

One of ~100 nodes in ComfyUI-ArchAi3d-Qwen by Amir Ferdos (ArchAi3d). ComfyUI Manager → search "ArchAi3d Qwen" → install → restart, or:

cd ComfyUI/custom_nodes
git clone https://github.com/amir84ferdos/ComfyUI-ArchAi3d-Qwen.git
cd ComfyUI-ArchAi3d-Qwen
pip install -r requirements.txt

Restart. It's a pure-torch utility - no deps beyond ComfyUI's own. Free for personal use; commercial use needs the paid license.

Notes

There are dozens of solid-color image nodes out there, so this one earns its place only if you're already in the ArchAi3d ecosystem (for the Qwen presets and the reference-image sizing). The preset_dimensions dropdown is the only thing specific to this pack - if a preset size looks odd (like 1664×832), that's deliberate, every entry is divisible by 32. Watch out for the hex input swallowing the # either way - the node accepts both forms, so a stray # isn't an error, just be consistent.