Nodes/comfyui-auto-anime-prompts/πŸ“· Passport Tile (4x6)
ComfyUI Node

πŸ“· Passport Tile (4x6)

Turn one passport photo into a printable 4Γ—6 sheet, no photo editor needed

By jluo-githubΒ·Created 7 months agoΒ·Updated 7 months agoΒ· 0
πŸ“· Passport Tile (4x6)
  • image
  • tiled_image

Passport Tile (4x6) is the last node in the passport pipeline of the comfyui-auto-anime-prompts pack, and it's the one that saves you from opening GIMP. It takes a single square passport photo and tiles it four times onto a standard 4Γ—6-inch print sheet at 300 DPI - 1800Γ—1200 pixels - so you can save one image and order a single 4Γ—6 print from any photo kiosk.

It's the print-shop trick automated: passport services charge a premium for a sheet that's four copies of the same photo, and the machine that makes them isn't magic. The node is the machine.

How it works

From nodes/passport_photo.py, it's a short, deterministic routine:

  1. Take your image (ideally the 600Γ—600 output of Passport Resize); if it isn't 600Γ—600 it gets resized to that first.
  2. Create a white 1800Γ—1200 canvas (4Γ—6 inches at 300 DPI).
  3. Paste the photo at four positions: a 2Γ—2 grid, centered horizontally with a 300px margin on each side.

Output: one tiled_image, ready to save and print.

There's exactly one input - image - and no settings to misconfigure, which is honestly the best feature a node like this can have. The layout is hard-coded the way kiosk layouts are: 2 across, 2 down, no fuss.

The gotcha, same as its sibling

Like Passport Resize, this node processes only image[0] - the first frame of the batch. It doesn't tile four different photos onto one sheet (a real need if you're printing four different people); it tiles the same photo four times. If you want a mixed sheet, you'd have to build it yourself. That limitation is exactly what the 2Γ—2-of-one-photo use case expects, so it's rarely a problem - just don't assume a batch of four becomes a family sheet.

Install

Standard for the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/jluo-github/comfyui-auto-anime-prompts

Then restart ComfyUI, or use Manager and search "Auto Anime Prompt". No extra Python dependencies (Pillow/numpy/torch are already in ComfyUI), no model downloads, MIT licensed. This node works standalone on any square-ish image - it doesn't need the Qwen editing half of the pipeline at all.

Full pipeline, for context

The intended flow, one node after another: Passport Prompt writes the edit instruction β†’ Qwen-Image-Edit executes it β†’ Passport Resize enforces the 2Γ—2 spec β†’ Passport Tile builds the printable sheet. Each node is small and honest about its single job. That's a nice bit of design for a pack that's otherwise all anime tags - and the tile node is the one you can reach for even if you never touch the rest.

Categoryimage/passport

Inputs (1)

NameTypeDefaultDescription
imageIMAGEβ€”

Outputs (1)

NameTypeDescription
tiled_imageIMAGEβ€”