Description This is an Anima workflow that generates high-quality Silly Tavern character expressions in just minutes. Gives you tons of options (portrait mode, background creator, low resolution, manual selection, additional prompt tags, etc.), yet, it's designed to be easy as just clicking "run" upon first load. In addition, it will generate sex expressions for all non-extreme positions, as well as combat poses. (Currently prototype mode, meaning this is highly unrefined, so expect a lot of updates) Update 0.5 Notes: Now the operation is starting to come together, as far as functions are concerned. As you can guess, multiple characters are now supported. Another big addition is making random expression generation optional. You now have the choice to generate each expression from start to finish. The nodes have been arranged to be more compact. Tweaking, spelling errors, fixes, that sort of thing. Look at the patch notes on the right if you want a better list. The last big thing is execution abort if you some how end up with a invalid expression number. (This version I'd consider more experimental with the number of big changes) InstallationPre-Requirements:<ul><li>[ <a target="_blank" rel="ugc" href="https://github.com/Comfy-Org/ComfyUI-Manager">Download</a> ]: ComfyUI Manager (It allows you to install missing nodes) (Or: pip install -U --pre comfyui-manager)</li><li>[ <a target="_blank" rel="ugc" href="https://gist.github.com/kevinjwesley-Collab/d550d30122a79c537ed104840bc1406c">Download</a> ]: Pixel Color Picker (It's not included in the ComfyUI Manager Repos) (Drop the <a target="_blank" rel="ugc" href="http://PixelColorPicker.py">PixelColorPicker.py</a> into your ComfyUI/custom_nodes folder)</li><li>[ <a target="_blank" rel="ugc" href="https://github.com/yolain/Comfyui-Easy-Use">Download</a> ]: Easy-Use Nodes (For wildcards and style selector)</li><li>Download the included ZIP file and extract it to your ComfyUI directory. (Contains wildcards and example character style sheet as well as the workflow) </li></ul>Operation<ul><li>Open ComfyUI, press 'W' for Workflows, click on ST_Expression_Gen </li><li>Then select a character from the Character Styles node. </li><li>Click 'Run' (Or batch run) </li><li>It will generate a random expression in this format: expression-##.avif (As well as a JSON containing the prompt and the seed) </li><li>Copy all avif files to SillyTavern/data/default-user/character/&lt;character&gt; </li><li>Restart SillyTavern, it now should recognize your character expressions</li></ul>Making your own character:<ul><li>Go to ComfyUI/custom_nodes/comfyui-easy-use/styles </li><li>Open the Characters.json (Or better yet, make your own, so your characters don't get overwritten on update) (Use Characters.json as a template.) (Use only character tags) (No quality tags, no action tags, gestures, posture, poses, camera, etc..) (Only use anatomy and clothing tags.)</li></ul>Since this is in prototype mode, you'll probably have to curate about 50% of the images and delete the ones that don't turn out well. QuestionsIs it perfect? No, but it's output will outdo 99% of other expressions out there. And, I have done my best to mitigate any limitations dealing with auto generation expressions. It's also in prototype phase, there are some wrinkles to iron out. What model should I use? Start with WAI - <a target="_blank" rel="ugc" href="https://civarchive.com/models/2544636/wai-anima">https://civarchive.com/models/2544636/wai-anima</a> How do I get Silly Tavern to use the extra expressions? Set expressions to Main API. Add custom expression for each one you want. (You only have to do this once.)

SillyTavern Auto Expressions

<ul><li>Multiple fixes (Background now works)</li><li>Tag updates to make it more coherent </li></ul>

Files with SHA256: 7ad3f84d6f89a6d19846dad1666b7828184565226d5d96c31372933f9c6c819f