Character Prompt Builder - Male Poses

Want your male character standing in a boxer's stance, kneeling in a combat-ready crouch, or sitting in a throne like he owns the room? Character Prompt Builder - Male Poses spells all of it out for you in full sentences - no more "man standing" and hoping. It's the male counterpart to the Female Poses node and the third link in the chain: Male Person → Male Fashion → Male Poses → Scene & Generate.

The critical gotcha is shared with its female sibling, so don't skip it: only one of the four pose categories is ever used. The node scans standing → kneeling → sitting → laying down and keeps the first one set to something other than "-", discarding the rest silently. Fill in two and you'll get one, and the node won't warn you. Pick a category, leave the others on "-".

What's here

• Poses - 26 standing options ("standing with hands in pockets", "standing in a boxer's stance", "standing at attention, military style"), 11 kneeling (proposal, prayer, combat-ready, "kneeling tying shoelace"), 16 sitting ("sitting in a throne-like position", "sitting back in chair, one ankle on opposite knee"), 11 laying down (face-down, on side propped on elbow, sprawling).
• props - 36 options with a distinctly male catalog: sword, gun, shield, hammer, axe, bow, briefcase, cigar, beer bottle. Plus props_color (48) to style them.
• Placement (optional) - the same hand_placement (11), leg_placement (10), and head_placement (8) refiners as the female node. These layer on top of your chosen pose rather than competing with it.
• custom_pose - free text for the pose the catalog doesn't have, merged verbatim into the final prompt.

How it works

Standard pack mechanism: settings_in dict in, pose choices stamped on, settings dict out. No text generation here - the Scene & Generate node is where the prose happens. Because the pose fields are already full sentences, whatever the Scene node writes reads like a direction to a model, not a tag list - which is exactly what current LLM-encoded text encoders want.

Installing

ComfyUI Manager → search Character Prompt Builder → Install → restart. Or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/euan-gwd/comfyui-character-prompt-builder

No pip dependencies, no model downloads - the pack runs on the Python standard library. The pose options live in character_prompt.json, fetched from GitHub on first use if the bundled copy is missing; offline, you get a thinner fallback list.

Troubleshooting

• Wrong pose in the image - almost always the "first non-dash wins" rule. Verify only one of the four pose categories is set, then re-run.
• Pose ignored entirely - confirm the chain actually passes through this node: Person → Fashion → Poses → Scene. Wiring Person straight to Scene skips the pose data.
• Prop missing - some checkpoints need the prop near the pose phrase to bind them; if "holding a sword" keeps vanishing, put a one-line clarification in the Scene node's prompt_suffix.