As you may know, I create wildcard packs regularly to test various LoRAs. This one is no different. In order to test MORE poses, I needed an expanded pack. This wildcard packs here are built from the images training data and contain a README.md that has info about use as well as a compare set of prompts for comparing yourself between danbooru and english style prompts. (If you are reading this before one or the other comes out... surprise 🎁 coming soon. <h2 id="more-expressive-poses-in-two-prompt-formats">More expressive poses, in two prompt formats</h2>These composable wildcard packs build prompts around movement, body position, and camera angles: reaching, twisting, balancing, reclining, flying, and interacting with the surroundings.Both draw from captions for the same training images, with wording suited to different prompting workflows:<ul><li><code>anima-pose</code> uses comma-separated Danbooru tags with spaces between words, intended for Anima. This is not an underscore-separated tag pack. (and that shift was one of my difficulties while testing and why I had to build this)</li><li><code>more-poses</code> uses descriptive English prose for models such as Krea2, with detailed and shorter prompt options.</li></ul><h3 id="start-here">Start here</h3>Place the folders inside your wildcard directory, preserving their names and subfolders. Use a ComfyUI wildcard node that supports nested/recursive expansion. In the ZIP file in the JSON of a workflow that has nodes that will work configured to work if you load those nodes. Anima — space-separated Danbooru tags<ul><li>Complete randomized prompt: <code>__anima-pose/pose__</code></li><li>Pose + camera for your own character: <code>__anima-pose/pose_only__</code></li><li>Rear-view composition: <code>__anima-pose/rear_pose__</code></li><li>Side-profile composition: <code>__anima-pose/profile_pose__</code></li></ul>English prose — for workflows such as Krea2<ul><li>Complete descriptive prompt: <code>__more-poses/pose__</code></li><li>Shorter complete prompt: <code>__more-poses/pose_short__</code></li><li>Pose + camera for your own character: <code>__more-poses/pose_only__</code></li><li>Rear-view composition: <code>__more-poses/rear_pose__</code></li><li>Side-profile composition: <code>__more-poses/profile_pose__</code></li></ul>Start with the complete prompt option, or append a pose-only call to your own character and outfit description. See each pack’s README.md for individual categories and more customization examples.<h3 id="what's-inside">What’s inside?</h3>Standing, seated, ground-level, movement, combat, acrobatics, and flight variations, plus activities such as playing instruments and photography. Supported poses use railings, branches, walls, stairs, and other surfaces.The packs keep related details together: pose and gaze, hands and props, wings and flight framing. Camera choices emphasize full-body views, low and elevated angles, perspective, foreshortening, and tilted compositions.You can also call individual categories, edit the clothing and appearance banks, or append an optional style. Eight matching comparison prompts let you try corresponding English and tag descriptions without random changes to the composition.A few useful notes: Add your own model triggers or LoRA settings as needed. Pose-only prompts may include props or supporting surroundings. Rear/profile calls already include camera instructions. When editing TXT files, leave no blank lines or trailing line ending after the final content line.See each pack’s README.md for the full category list, customization examples, and usage details. The compare/README.md explains the matched comparison prompts.

More Poses Wildcard Pack

English version, will work with any English translating diffusion model. The embedded JSON is configured for Krea 2 only.

Files with SHA256: 89aac96b0cbb99e208710e3471047f3f92413bf4ac4e00f3e1defe426f5b8e86