EmAySee Twenty Float To Text V2

EmAySee Twenty Float To Text V2 is the smarter sibling of the pack's plain Twenty Float To Text And Float. Same idea - twenty float sliders describing body attributes, with 0 meaning "don't mention it" - but V2 stops forcing you to take the results as one big blob. It gives you the combined string and each attribute formatted on its own, so you can place them wherever you want in a prompt.

Here's the shape of it. Twenty float inputs, all defaulting to 0 with a -2 to +3 range in 0.01 steps: Breast Size, Breast Sag, Areola Size, Puffies, Belly Fat, Waist Size, Butt Size, Thigh Size, Thigh Spacing, Body Type, Pregnant, PHP, and eight generic Extra 01–08. (Yes, "PHP" - the author's shorthand for pubic-hair-peek. And yes, they fixed the "Areloas" typo from the first version.) Non-zero values are formatted as Label: 1.25 with two decimals.

The outputs are where V2 earns its name:

• combined_text - the newline-joined string of every non-zero attribute, same as the older node.
• T_ outputs* - twenty separate STRING outputs, one per attribute. Each holds just its own Label: value line, or an empty string if that attribute is 0.
• F_ outputs* - twenty FLOAT pass-throughs, echoing the raw slider values back out for wiring into LoRA strengths and the like.

Why bother with the individual T_ strings? Because combined_text is all-or-nothing. If you want "Breast Size: 1.5" in the positive prompt but only "Belly Fat" and "Waist Size" in some other field, the blob version forces you to post-process the text somewhere else. With V2 you concatenate just the T_ outputs you want - and because zero-valued attributes come out as empty strings, you can chain them together without filtering anything. It's a small thing, but in a slider-driven workflow it's the difference between "works" and "works without a text-cleanup node bolted on."

Mechanically it's the same zero-dependency story as the rest of the pack: a Python loop, no models, no API, no network. (It does import torch at the top and never uses it - harmless, just a tell that this is hand-rolled.) It lives in the pack's EmAySee/Text category.

How to install

Same install as every node in this pack - install the pack once:

cd ComfyUI/custom_nodes
git clone https://github.com/EmAySee/ComfyUI_EmAySee_CustomNodes

Restart ComfyUI afterward, or search ComfyUI_EmAySee_CustomNodes in ComfyUI Manager and let it do the work. No requirements.txt, no model files to download, and nothing heavy - just ComfyUI's own Python environment. You'll know it loaded when the console prints the "SPECTRE v5.0" banner.

Where people trip up

• 0 is "absent," not "zero." Every attribute that sits at 0 gets dropped from combined_text and becomes an empty string in its T_ output. Drag a slider back to 0 and that attribute quietly disappears from your prompt.
• The per-attribute outputs are strings, not numbers. If you're used to grabbing values off a slider node, remember the T_ outputs are formatted text (Breast Size: 1.50) and the F_ outputs are the numbers.
• Don't mix it with the first version's workflow. The label set changed (Puffies vs Puffie, PHP vs Pubic Hair Peek), so a saved workflow wired for the original will show disconnected inputs after a swap.

If all you need is a single combined prompt line, the original node is smaller and does the job. The moment you start routing different attributes to different parts of a prompt - which is the whole point of attribute-slider work - V2 is the one you'll actually reach for.