Icy ConditioningCombine

IcyConditioningCombine is the IcyHider clone of ConditioningCombine, and it's the node that answers "can I run two prompts at once?" Yes - and this is how the sampler hears both.

What it does

Two conditioning inputs, one output:

• conditioning_1 and conditioning_2 - two CONDITIONING streams (positive branches from separate CLIPTextEncode nodes, or the outputs of regional/ControlNet-modified conditionings).
• Output: a single CONDITIONING that carries both.

Mechanically, combine is a list concatenation: the two conditionings sit side by side in one list, and the sampler applies each one independently. The output isn't a blended middle - it's "prompt A, and also prompt B," each steering the denoise with its own strength and its own optional area metadata intact.

Why that's useful

• Multiple positive prompts. Encode two separate prompts (say, "studio portrait, soft light" and "green velvet jacket") and combine them, and the model has both to work with - useful when one long prompt gets tangled or when you want to build prompts from parts.
• The regional pattern. When you set up region prompting, each region's conditioning gets combined back together so the sampler honors all regions at once. ConditioningCombine is the merge point.
• Prompt building blocks. Keep reusable prompt fragments as separate encode nodes and combine them per-generation instead of retyping.

The sibling you should know: ConditioningConcat glues two prompts' tokens into a single sequence (different mechanism, similar-looking node). Combine = "use both." Concat = "treat as one run-on prompt." For most "I want both prompts" cases, combine is what you want.

Where people get burned: combining a positive and a negative conditioning and expecting the sampler to know which is which. Combine doesn't label its inputs - the sampler still treats the merged list as "the positive branch." Keep negatives on their own path to the sampler's negative input.

The "Icy" caveat

Outputs CONDITIONING, no image preview, so the Icy wrapper is cosmetic relabeling - the pack's clone of every core node. The list-concatenation math is exactly core ConditioningCombine.

Install

ComfyUI Manager → search "IcyHider" → install → restart, or:

cd ComfyUI/custom_nodes
git clone https://github.com/icekiub-ai/ComfyUI-IcyHider.git

Restart. No Python deps, no models bundled. Node under IcyHider Comfy Core. If combining "does nothing," double-check that both inputs are real positive-branch conditionings - an empty or zeroed branch combines as empty, and the output is only as strong as its weakest input.