Image Tiled

A naming note first, honestly: WAS's own README lists a node called "Image Tile" (singular) - "Split a image up into a image batch of tiles. Can be used with Tensor Batch to Image to select a individual tile from the batch" - and comfy.icu's catalog surfaces this one as "Image Tiled." I'm treating these as the same node given how close the names and jobs are, but flagging the mismatch rather than pretending it's a confirmed 1:1 match.

What it does

It cuts one image into a grid of smaller pieces and returns them as an image batch rather than one image - the same tensor-batch shape ComfyUI uses everywhere else a "batch" comes up. From there, the README points at Tensor Batch to Image as the natural next step: pull a single tile out of the batch by index so you can look at or process one piece at a time.

Why you'd reach for it

Tiling is the backbone of a lot of high-resolution workflows - processing a huge image tile-by-tile (upscaling, detail passes, tiled diffusion techniques) keeps memory and compute manageable in a way that running the whole image through a heavy operation at once doesn't. This node is WAS's manual entry point into that pattern: cut it up yourself, process the pieces however you want, and handle reassembly on your own.

The inputs and outputs that matter

An image in, a tile size or grid configuration (the exact widget names I can't confirm without a schema), and an image batch of tiles out. Pair the output with Tensor Batch to Image if you need to isolate and work on one tile at a time rather than processing the whole batch uniformly.

How to install it

Bundled with the full WAS Node Suite:

• ComfyUI Manager - search "WAS Node Suite", install, restart.
• Manual - cd ComfyUI/custom_nodes && git clone https://github.com/WASasquatch/was-node-suite-comfyui, install requirements.txt (portable: python_embeded\python.exe -s -m pip install -r requirements.txt), restart.

No model download - pure image-splitting logic.

Common issues & troubleshooting

Uneven tiles at the edges. If your image's dimensions don't divide cleanly by your tile size or count, expect the last row or column of tiles to come out a different size, or the image to get padded to fit - a normal consequence of tiling math, not a bug specific to this node. If your downstream processing assumes uniform tile sizes, check the edges before trusting the whole batch is consistent.

No built-in reassembly step. This node splits; it doesn't stitch back together. Once you've processed your tiles individually, putting the image back into one piece is on you to handle with whatever compositing approach fits your workflow - this pack's own image-stitching and paste-crop nodes are worth a look if you need that step.

Standard pack caveat: WAS Node Suite has been unmaintained since December 2023, and the recurring issue is the whole suite failing to import after a ComfyUI update from a stale pinned dependency. Reinstall requirements.txt against the right Python environment if that hits you.