Danbooru tags
Danbooru tags are the structured tagging vocabulary inherited from the Danbooru imageboard — character, style, scene, expression tags — used heavily as the prompt language for anime-style diffusion models in 2026.
Anime-tuned diffusion models (NovelAI, Pony Diffusion, Animagine, AnythingV variants) are trained on captions in Danbooru-style tag format. Prompts in those models work best when they look like a Danbooru search query: comma-separated tags grouped by category (artist, character, copyright, general). The structure gives precise control — "1girl, blonde hair, ahoge, blue eyes, school uniform, masterpiece, best quality" is a typical anime prompt. Generic photoreal models (Flux, GPT-Image, Imagen) do not respond to Danbooru tags the same way; they expect natural language. Mixing styles in one prompt usually degrades both.
When to use danbooru tags
- Prompts for anime-tuned diffusion models.
- Character / scene control via tag composition.
Common mistakes
- Mixing Danbooru tags with photoreal-model natural-language prompts.
- Overloading tag count past ~80 — models start losing focus.
FAQ
What is danbooru tags?
Danbooru tags are the structured tagging vocabulary inherited from the Danbooru imageboard — character, style, scene, expression tags — used heavily as the prompt language for anime-style diffusion models in 2026.
When should I use danbooru tags?
Prompts for anime-tuned diffusion models. Character / scene control via tag composition.
What are the most common mistakes with danbooru tags?
Mixing Danbooru tags with photoreal-model natural-language prompts. Overloading tag count past ~80 — models start losing focus.
Related terms
- Diffusion model — A diffusion model is a generative neural network that creates images, video, or audio by iteratively denoising random noise toward a learned target distribution.
- Negative prompt — A negative prompt is text that tells an image, video, or audio generator what to avoid producing — the opposite of the main prompt.
- LoRA (Low-Rank Adaptation) — LoRA is a fine-tuning method that trains a small set of low-rank adapter weights on top of a frozen base model — cheaper to train and store than full fine-tuning.
Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/danbooru-tags.md.