GR Image Multiplier

This is the one that actually does what "multiplier" implies: feed it images, tell it how many copies you want, and it hands back a batch that's your input repeated that many times over. It's a batch-padding node - useful anywhere you need N copies of the same image(s) queued up for something downstream that varies per-item, like a batch of seeds or a batch of LoRA weights.

How it works

You give it an images batch and an integer multiplier. It repeats the batch that many times to build the output. Where it gets useful is the two boolean switches: interleave and random_order. If your input is a batch of several distinct images and you multiply by, say, 3, plain repetition would give you [A, B, C, A, B, C, A, B, C]. Turn interleave on and you get [A, A, A, B, B, B, C, C, C] instead - grouping matters if whatever consumes the batch downstream processes it in chunks tied to the original image count. random_order shuffles the final sequence, seeded by the seed input so it's reproducible - handy when you specifically don't want the multiplied batch coming out in a predictable block pattern (say, you're feeding a grid preview and want variety visually spread across it rather than clustered).

The inputs and outputs that matter

• images (IMAGE, required) - your source batch, one or many images.
• multiplier (INT, default 1, range 1–10000) - how many times to repeat the batch. Start small; 10,000× a batch is a VRAM bill you don't want to accidentally trigger.
• interleave (BOOLEAN, default off) - group repeats by original image vs. by full-batch cycles. Flip it on if the order in which items appear matters for what's downstream.
• random_order (BOOLEAN, default off) - shuffle the final output order.
• seed (INT) - only matters when random_order is on; same seed gives you the same shuffle every time.
• IMAGE (output) - the multiplied batch, ready to feed a sampler, a save node, or anything else that wants more items than you started with.

How to install it

Via ComfyUI Manager: search "GraftingRayman", install, restart. Manually:

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

then restart ComfyUI. Like every node in this pack, it needs OpenAI's CLIP package installed on top of ComfyUI's own CLIP handling, or the whole pack fails to import:

# portable build
.\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

Out of memory after multiplying. This is the obvious footgun with a multiplier that goes up to 10,000 - every copy is a real image in VRAM/RAM until something consumes the batch. If you're testing, start at 2–4x and work up rather than guessing at a big number and waiting for the crash.

Node doesn't appear in search. Same as the rest of the pack: check the OpenAI CLIP install above before anything else. It's a one-time dependency failure that takes out every GraftingRayman node, not something specific to this one.

Downstream node treats the batch oddly. If a node further down the graph expects images grouped a particular way (per-item vs. per-cycle), that's what interleave is for - flip it and see if the behavior matches what you expected. It's the most commonly missed toggle here because the default (off) is the less intuitive grouping for most use cases.