ComfyUI-GridSplit
Two complementary grid tools for ComfyUI — no model, no weights, CPU, any resolution:
- Auto Grid / Carousel Split — detect how a stitched image is divided and split it into its panels.
- Grid Stitch — the inverse: repeat one image into an R×C grid, scaled to a target megapixels.
Auto Grid / Carousel Split
Auto-detect how a stitched image is divided and split it into its individual panels. Handles three kinds of layout automatically:
| Layout | example | how |
|---|---|---|
| Regular grid | 2×2 of variations | global seam projection |
| Uneven grid | 3 side-by-side, unequal widths | global seam projection |
| Irregular collage ("bento") | different split lines per region, a cell split again | recursive guillotine on the gutters |
It auto-routes: if the image has solid separator bands (black/white/any flat gutter) it uses recursive guillotine decomposition (each region finds its own local cuts, so irregular layouts work). If the panels are seamless (abut with no gutter) it uses global row×column seam detection. Content edges (faces, text, etc.) never trigger a split.
Validated: 2048² 2×2 → 4 tiles · 2728×1536 uneven 1×3 → 928/976/824 tiles · 3344×1880 irregular 10-panel collage (incl. a column that's itself split in two).
Examples
Simple 2×2 grid → 4 clean panels (bottom-right shows the preview output with detected boxes in red):
Irregular collage → 10 panels — note the bottom-middle cell is itself split into two, while the pool and chalkboards stay whole:
Inputs / outputs
Inputs
| name | what |
|---|---|
image |
the stitched grid / carousel / collage |
sensitivity |
default 1 (standard). Higher = splits more eagerly / catches weaker seamless seams. Only affects the seamless path. |
min_panel_px |
reject any split that would make a panel smaller than this |
edge_trim |
shave N px off each panel edge (kills blend / halo / anti-aliased gutter fringe) |
Outputs
| name | what |
|---|---|
panels |
list of images (tiles keep their native size) → feed into SaveImage to write N files |
count |
number of panels detected |
preview |
the input with detected panel boxes drawn in red — for visual QA |
panels is a list because panels have different sizes and can't share one batch tensor.
A non-grid image passes through unchanged as a single panel.
Grid Stitch
The inverse of the splitter: take one image you like and repeat it into an R×C grid, scaled so the whole thing hits a target megapixel count. Cells preserve the image's aspect ratio (no distortion) and tile seamlessly.
Inputs
| name | what |
|---|---|
image |
the image to tile (a batch of N fills the cells cyclically; a single image just repeats) |
rows, cols |
grid dimensions, e.g. 3×3 or 3×4 |
megapixels |
target total size of the stitched grid |
scale_method |
resampling filter — same set as ComfyUI's Upscale Image (nearest-exact / bilinear / area / bicubic / lanczos), default bicubic |
Outputs
| name | what |
|---|---|
grid |
the stitched grid image |
width, height |
final output dimensions |
Example: a 1448² image at 3×4 @ 4 MP → a 2308×1731 grid, each cell ~577×577, no distortion.
Install
cd ComfyUI/custom_nodes
git clone https://github.com/workordie/ComfyUI-GridSplit.git
# restart ComfyUI
No dependencies beyond torch/numpy (already in ComfyUI).
Example workflow
Drag example_workflows/autogrid_example.json
onto the ComfyUI canvas — it generates an image and splits it, wiring panels
into SaveImage and preview into PreviewImage.