XY Input: Manual XY Entry

Every other XY Input node is specialized - one for samplers, one for clip skip, one for LoRAs. Manual XY Entry is the escape hatch. You pick which parameter to vary, then type the exact values you want by hand. It's the node you reach for when you know precisely which values to test and don't want the other nodes' range-and-count machinery deciding for you.

Like the rest of the XY family, it does nothing on its own - it feeds an XY Plot node.

Why you'd reach for it

Say you want to compare CFG at exactly 3, 5, and 8 - not a range that gets divided into three, but those three specific numbers. Or steps at 8, 20, and 40. Manual entry lets you type the values straight in, which is faster and more precise than fiddling with first/last/count when the set you care about isn't evenly spaced. It's also the way to plot a parameter that doesn't have its own dedicated XY Input node.

The inputs and outputs

• plot_type - what to vary. The list is broad: Nothing, Seeds++ Batch, Steps, StartStep, EndStep, CFG Scale, Sampler, Scheduler, Denoise, VAE, Positive Prompt S/R, Negative Prompt S/R, and a few more. This tells the plot which knob your typed values control.
• plot_value - a multiline text box where you enter the values, one per line (or in the format that plot type expects). This is the whole point of the node: you write the exact axis by hand.

The output is labelled X or Y (XY type). Wire it into the X or Y input of an XY Plot node.

Installing it

ComfyUI Manager → search Efficiency Nodes for ComfyUI, or:

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

Restart after. No extra downloads.

Common issues

The main friction is format. Your plot_value entries have to make sense for the plot_type you picked - numbers for CFG or Steps, valid sampler names for Sampler, and so on. A typo'd sampler name or a stray character in a number line will fail that cell or the plot. When in doubt, keep it to one value per line and match exactly what the corresponding dropdown would show elsewhere.

Nothing as a plot type is a valid "leave this axis empty" choice - handy if you only want to vary one axis.

And it only functions through an XY Plot driving an Efficient KSampler; on its own or into a stock sampler it does nothing. Pack-wide: a missing node usually means IMPORT FAILED at startup (often a pip freeze error) - update ComfyUI, the packages and the node.