GR Tile and Border Image Random Flip

Identical schema to its sibling GR Tile and Border Image - same image, columns, rows, border, colour - but the behavior underneath is different: instead of repeating your source image identically into every grid cell, each tile gets randomly mirrored. That's the classic trick for breaking up the obvious "this is one image copy-pasted across a grid" look you get from straight tiling, useful for quick faux-seamless textures or patterns where an exact repeat would read as too mechanical.

How it works

columns and rows set the grid, same as the non-random sibling; border and colour add an optional divider between tiles. The flip is applied per-tile and randomly, as the name says, with no exposed control over which tiles get flipped or how.

Be clear-eyed about the real limitation here: there's no seed input anywhere in this node's schema. That means the randomness isn't reproducible - run the exact same inputs twice and you'll get two different flip patterns, with no way from this node alone to lock in a specific result you liked and get it back later. It's a one-shot "roll the dice and see what you get" node, by omission if not by explicit design. If reproducibility matters to your workflow, save the output image itself when you land on a result you want, rather than trying to regenerate it from the same graph.

The inputs and outputs that matter

• image - the source to tile and randomly mirror.
• columns / rows - grid dimensions, both minimum 1.
• border / colour - optional divider between tiles.

One output: IMAGE, the tiled-and-randomly-flipped composite.

How to install it

ComfyUI Manager, search GraftingRayman. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/GraftingRayman/ComfyUI_GraftingRayman

Restart afterward. The pack fails to import entirely without OpenAI's CLIP installed separately, even though this node has nothing to do with CLIP - portable ComfyUI: .\python_embeded\python.exe -m pip install git+https://github.com/openai/CLIP.git; system Python: pip install git+https://github.com/openai/CLIP.git.

Common issues & troubleshooting

Pack fails to load. Check your startup console for the CLIP import error and run the pip install above.

Can't get the exact same result twice. Expected - there's no seed on this node, so the flip pattern is different every run by design (or at least by omission). Save the output you like rather than trying to reproduce it from the graph.

Run is slow or memory-heavy on a large grid. Same as its sibling - columns and rows have no ceiling in the schema, so canvas size scales directly with tile count. Start small.

Want a consistent, non-random tiled result instead. That's the plain GR Tile and Border Image - same schema, no randomness, exact repetition of the source.