Annotation Labels

Annotation Labels is the simplest node in the Lukutar prompt family, and it's honest about it: it takes the ANNOTATIONS output from Prompt Annotate and renders the set of region labels as one comma-separated string. body, face, hair, background - that's it. No parsing, no transforms, no dark corners.

Which sounds pointless until you actually need it. The realistic uses are exactly what the README names: feeding a downstream label picker, or logging which regions a workflow covers so you can confirm your annotation and your detector agree before you burn a render on a mismatched setup. When you're debugging a regional workflow - "why did the hair region get nothing?" - a quick glance at the labels list tells you whether the annotation even has a hair region or whether you typo'd the tag in Prompt Annotate.

Inputs and output

• annotations - the ANNOTATIONS output from Prompt Annotate. Strictly typed, as with the rest of the family: wire in anything else and it raises a clear error.
• include_common (default off) - whether to include the implicit all label that the unmarked common part of the prompt carries. Leave it off and you get just the tagged regions; flip it on if you want all in the list too.

Output is one labels (STRING), comma-separated, in first-appearance order with duplicates removed. Two small edge cases worth knowing: with include_common on, the implicit all label of the common part is included; with no tagged regions and include_common off, the output is an empty string. Neither is a failure - they're just the node telling you the truth about your prompt.

Installation

Part of ComfyUI-LukutarNodes (MIT). ComfyUI Manager: Custom Nodes Manager → Install via git URL → https://github.com/Mistress-Lukutar/ComfyUI-LukutarNodes, or:

cd <ComfyUI>/custom_nodes
git clone https://github.com/Mistress-Lukutar/ComfyUI-LukutarNodes

Restart ComfyUI. Only extra runtime dependency is opencv-python (>= 4.8) - Manager installs requirements.txt; manual installs may need pip install opencv-python into ComfyUI's python. No models, no keys. Category Lukutar/Prompt.

When to bother

Honestly, this is a node you add when you're building a workflow, not one that changes results. Reach for it when you want to see the region set at a glance (wire it to a Show Text node), or when you're automating something that picks labels. If you just want one label's text for a detail pass, Annotation Segment is the node you want instead - this one is the overview, that one is the extractor. Both are thin, both are cheap, and neither will surprise you.