Anima Pose Tag Selector

Here's the situation Anima users live with: pose conditioning is the model's biggest gap. The only ControlNet-ish option is a community LLLite sample weight that its own author calls deliberately weak, and the pose model in particular is "noticeably weaker control than the others." If your workflow needs reliable pose, Illustrious is still the answer - but if you're staying on Anima, your practical tool is tags. The Anima Pose Tag Selector gives you a curated library of pose prompt tags to reach for.

It's part of the newer wave of Anima-Tools nodes (the README still doesn't mention background or pose selectors - they shipped after it was written; the code is the source of truth). You'll find it under AnimaArt as "Anima Pose Tag Selector", a visual 2:3 preview grid of poses you click into the prompt.

Mechanically it's the same thin cleaner as its siblings: split pose_tags on commas, trim, drop empties, join - no @ rewriting, and a _raw_: prefix passes a tag through untouched. The value isn't the Python; it's that the bundled pose database contains tags the model actually understands, so you're not guessing whether "crouching" or "sitting on a chair" or the specific booru phrasing is the right one.

The inputs that matter

• pose_tags - the comma-separated pose list the panel fills.
• mode - append (default) puts poses ahead of opt_prompt; override replaces it.
• opt_prompt - optional existing prompt to merge with.

Single text STRING output (with the pack's trailing ", "), which feeds a CLIPTextEncode positive or the pack's Anima Prompt nodes' pose_tags input.

Install

Zero-dependency pack, same install as everything else:

cd ComfyUI/custom_nodes
git clone https://github.com/nregret/Comfyui-Anima-Tools

Or ComfyUI Manager → search "Anima Tools" → Install, then restart. No pip, no model files; the pose database is bundled, previews come from a CDN so the grid needs internet.

Be realistic about what tags buy you. A well-chosen pose tag steers the composition, but it won't pin a character to a precise skeleton the way a working ControlNet would. If pose accuracy is make-or-break, that's the one honest reason to keep an Illustrious setup around. If you're just tired of "how do I say this pose" friction, this node kills that problem stone dead.