XY Input: Seeds++ Batch

The simplest node in the whole XY Plot family, and one of the most useful. XY Input: Seeds++ Batch generates a batch of sequential seeds and lays their outputs across a plot axis, so you can see several variations of the same prompt at once. It's the "give me options" node - the fastest way to judge whether a prompt is consistently good or just got lucky on one seed.

How it works

There's exactly one widget: batch_count (default 3). The node produces that many seeds, starting from the seed set on your Efficient KSampler and incrementing, and outputs them as a single X or Y connection. The XY Plot node then renders one cell per seed.

That's it. No range, no target parameter, no modes. You pick how many variations you want and it makes them.

Why you'd use it over a normal batch

You could just raise the batch size on a normal latent and get multiple images. The difference is that Seeds++ Batch slots into the XY Plot grid, which means you can put it on one axis and sweep something else on the other - CFG, steps, sampler, checkpoint. Now you've got a matrix: "these five seeds across these four CFG values," all labeled and tiled in one execution. That's the thing a plain batch can't give you.

Used alone on a single axis, it's still the cleanest way to sanity-check a prompt: if four of five seeds look great, the prompt is solid; if one is great and four are a mess, you got lucky and should keep tuning.

The inputs and outputs

• batch_count (required, default 3, up to 50) - how many seeds to generate.

Output is X or Y (type XY), which goes to the XY Plot node's X or Y input. The XY Plot node plugs into an Efficient KSampler's script input, and the seed sequence starts from that sampler's own seed.

Installing it

ComfyUI Manager → search Efficiency Nodes for ComfyUI, install, restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/jags111/efficiency-nodes-comfyui

Restart ComfyUI. Nothing else to download.

Common issues

The universal XY Input catch is the only real one here: on its own the node does nothing visible. Its output has to run through the XY Plot node into one of the pack's Efficient KSamplers - a vanilla KSampler has no script input. If your seed grid never appears, that's almost always the reason.

Beyond that, be reasonable with batch_count. Each seed is a full generation, so a count of 50 on one axis crossed with a five-value sweep on the other is 250 images - that's a long run and a lot of VRAM for the tiled result. Start small.

Pack-wide, if Efficiency Nodes fails to import (IMPORT FAILED at startup, typically a pip freeze non-zero exit), this node vanishes with the pack. Update ComfyUI, update the node, and confirm your Python environment is healthy.