AHEKOT/ComfyUI_VNCCS

What it solves

Making a unique character that looks consistent across many images is much harder than generating a single picture. VNCCS solves this by providing a step-by-step workflow to define a character once and then generate sprites with controlled poses, outfits, facial expressions, and transparent backgrounds, all while keeping the character's appearance stable.

How it works

VNCCS is a collection of custom nodes and workflow JSON files for ComfyUI. It splits the process into stages:

• Character Creator: Set up the base character using text descriptions, presets for race, skin tone, body type, hair, etc., or use the Character Wizard to auto-fill options. You can generate previews and apply artistic styles from a built-in library.
• Character Cloner: Load an existing image (screenshot, drawing, etc.) and turn it into a workable character by auto-captioning and extracting features.
• Pose Studio: Choose or import poses, adjust body proportions, and generate sprites in those poses. Includes background removal with chroma key presets and optional SAM3 detail recovery.
• Clothes Designer: Create multiple outfits per character, either by describing clothing elements or using the Clothes Wizard, and clone clothes from images.
• Emotion Studio: Add facial expressions to sprites across chosen costumes, using models like Qwen Image 2.1 for face-replacement or Illustrious/Anima for full re-generation, with a denoise control to balance expression strength vs. character fidelity. The suite integrates model downloading via a Control Center and supports saving/versioning characters, styles, and costumes through API endpoints with revision conflict detection.

Who it's for

VNCCS is aimed at visual novel developers, game artists, and content creators who need a reliable way to produce consistent character sprites for games or projects. It requires some familiarity with ComfyUI, but the provided workflows and easy-install option lower the barrier for newcomers.

Highlights

• Consistency across all images: A single character definition ensures the same face, body, and style appear in every sprite.
• Full pipeline: Covers character creation, cloning, pose generation, clothing design, and emotion application in one suite.
• Multiple model support: Switches between Illustrious, Anima, Qwen Image 2.1, Flux Klein9b, and MiniMax H3 depending on the task (character base, poses, emotions).
• Style library with previews: Built-in and user-defined drawing styles can be tested instantly via the Generate Preview button.
• Background removal built-in: Chroma key presets and optional SAM3 recovery help create clean transparent sprites.
• Workflow automation: Wizards for characters and clothes, plus auto-generated previews, reduce manual setup.