Optional Conditioning Merge
Combine 1-3 conditionings without rewiring for each
- cond1
- cond2
- cond3
- CONDITIONING
A small annoyance in stock ComfyUI: combining conditioning is usually a fixed two-input node, so if you sometimes want to merge two prompts and sometimes three, you end up either always wiring three inputs (with a dummy on the third) or swapping nodes depending on the job. This one sidesteps that by making every input optional and adapting to however many you actually connect.
What it does
There are no required inputs at all - cond1, cond2, and cond3 are all optional CONDITIONING slots. Connect just one and it effectively passes through. Connect two and it averages them evenly - each contributing 50% of the blend weight. Connect three and each contributes roughly a third. The node calculates the blend weight itself based on how many connections are actually live, rather than making you set a weight manually, and it layers the conditionings element-wise rather than just concatenating them. The single CONDITIONING output wires into a KSampler's positive or negative slot exactly like any conditioning would.
The practical use case: you've got a style prompt, a subject prompt, and maybe a mood/lighting prompt encoded separately (useful if you're iterating on each independently), and you want them merged into one conditioning without hardcoding "always exactly two" or "always exactly three" into your graph.
Installing it
ComfyUI Manager: search DemonAlone-nodes-ComfyUI, install, restart. Manually:
cd ComfyUI/custom_nodes
git clone https://github.com/DemonAlone/SimpeStringGenerator_ComfyUI
No dependencies or downloads beyond core ComfyUI.
Where people get burned
Leave all three inputs disconnected and you've got nothing to output - make sure at least one is wired before this feeds a sampler, or you'll get an error (or a silently broken generation) further down the graph.
Also worth being clear-eyed about what this isn't: it's an even, element-wise average, not regional or masked conditioning. If what you actually want is different prompts controlling different areas of the image - a character prompt on the left half, a background prompt on the right - this isn't that tool. This one blends everything into a single global conditioning; reach for a dedicated regional-prompting setup if you need spatial control instead of a blend.
Inputs (3)
| Name | Type | Default | Description |
|---|---|---|---|
| cond1opt | CONDITIONING | — | |
| cond2opt | CONDITIONING | — | |
| cond3opt | CONDITIONING | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| CONDITIONING | CONDITIONING | — |