Divide Image

This is the simplest node in the Divide and Conquer suite, and it does exactly what the name says: it takes an image and cuts it into tiles. No position picker, no single-tile mode, no decisions. You give it the image and the plan, it hands you every tile as a list, done.

The plan comes from the Divide and Conquer Algorithm node as dac_data - the same payload that carries tile size, overlap, grid, and processing order to every other node in the suite. That's the whole trick of the pack: the algorithm decides the geometry once, and every downstream node just obeys the plan, so the splitter and the combiner can never disagree about where a tile goes.

Inputs and outputs

• image - the image to split (usually the upscaled output of the algorithm node).
• dac_data - from the Divide and Conquer Algorithm node.
• IMAGE - all tiles as a list, ready to feed into per-tile processing.

That's genuinely the whole node. If you want to split and pick a specific tile, or pass just one tile through for testing, you want the sibling node Divide Image and Select Tile - which does everything this does and adds a position selector. This node is the version that always passes the full set through.

Why it exists

This is the older, plainer incarnation of the splitter, the one that predates the 2.0 rework - the changelog for 2.0.0 mentions being able to "generate all tiles or a single tile without fiddling with the links," which is exactly the freedom the newer node added. If this is the class that shows up in your workflow or in an old saved graph, it still works fine: it splits, you process, and Combine Tiles stitches the results back up using the same dac_data. On a current install you'll more likely find the newer node in the menus, but the old class name still resolves - comfy.icu keeps both.

Install

It ships with the rest of the ComfyUI Steudio pack - one install covers every node in the suite:

cd ComfyUI/custom_nodes
git clone https://github.com/Steudio/ComfyUI_Steudio

Restart ComfyUI after cloning, or install via ComfyUI Manager by searching ComfyUI Steudio. No pip dependencies, no model files to download for the nodes themselves. If a saved workflow references this node and it isn't on your machine yet, installing the pack is all you need.