Core Feature Overview
- Diffusion V5 Full and Curated: V5 Full and Curated are NovelAI-designed image models with different creative priorities; choose between experimentation and more curated anime illustration rather than assuming both renders behave identically.
- Natural Language and Tag Prompts: Describe a character in ordinary English or use tightly organized tags for clothes, poses and artwork details; compare prompt interpretations before committing to an established character design.
- Multi-Character Composition: V5 includes reworked positioning controls and supports complex scenes with multiple identified characters; panel layouts and overlapping bodies still deserve careful proofing.
- Anime and Furry Style Selection: The image workspace is built around illustrated character genres rather than default commercial photography; select the model and prompt style for the intended linework and character type.
- Inpainting and Image Refinement: Retouch a local area instead of rebuilding the entire composition; evaluate whether the revised facial expression, costume or hand remains consistent with surrounding line art.
- Enhance and Upscale Workflow: Use enhancement after choosing the strongest composition, especially when preparing prints or detailed comic crops; newly hallucinated textures can alter intentionally simple illustration.


