Your prompt
Organized prompt
By category
Paste a prompt above to sort its tags into categories.
A tool that tidies a messy Stable Diffusion prompt (comma-separated tags) the moment you paste it. Each tag is classified with a built-in dictionary into categories: quality (masterpiece, best quality, score_9…), subject & count (1girl, solo…), appearance & face (long hair, blue eyes, smile…), clothing (school uniform, thighhighs…), pose & action (sitting, looking at viewer…), composition & view (from above, full body…), lighting (cinematic lighting, backlight…), background & setting (forest, night…), style & artist (watercolor, anime, by …), and other. The output is re-joined in the order most SD users follow (quality → subject → appearance → … → style), so you can copy it straight into the prompt field of WebUI / Forge / ComfyUI / NovelAI. If a tag carries emphasis or weight syntax like `(masterpiece:1.2)`, the symbols are stripped only for classification while the output keeps exactly what you wrote. You can toggle removing duplicate tags and sorting each group alphabetically. It shows the tag count, group count, and a per-category breakdown, so you can spot what's missing (no lighting or background, for example). The dictionary is fixed strings bundled with the tool — no external API or translation service — so the prompt you paste is never sent anywhere. It all runs in your browser.
How to use
- Paste your prompt (comma-separated tags) into the "Your prompt" box.
- Tags are sorted into categories — quality, subject, clothing, lighting, style, and so on — and re-listed in the conventional order in "Organized prompt".
- Optionally toggle "Remove duplicates" and "Sort A→Z in each group", then Copy the organized prompt into your tool.
FAQ
Is the prompt I paste sent to a server?
No. Tags are classified in your browser using a fixed dictionary bundled with the tool, with no external API or translation service. The prompt you paste and the organized result are never uploaded or sent anywhere — everything is processed only on your device.
What order does it sort into?
Quality → subject & count → appearance & face → clothing → pose & action → composition & view → lighting → background & setting → style & artist → other, the conventional Stable Diffusion sequence. Order within each category is preserved, or you can switch to alphabetical with "Sort A→Z in each group".
What happens to weighted tags like (masterpiece:1.2)?
The parentheses and `:1.2` are stripped only when deciding the category (so it classifies on masterpiece), while the output keeps exactly what you wrote, weights and all. It reorders without breaking your weights.
Some tags aren't classified well.
Words not in the dictionary, or unique proper nouns, land in "Other". A misclassification only changes ordering — no tag is lost — so you can hand-edit the output box before pasting. The dictionary prioritizes the common, well-known tags.
Can it remove duplicate tags?
Yes. With "Remove duplicates" on, tags with the same content (treated as identical ignoring weight parentheses) keep only the first occurrence. Turn it off to reorder every tag as-is.
Does it work with WebUI / Forge / ComfyUI / NovelAI?
Yes. The organized output is a plain comma-separated prompt, so any tool that takes a text prompt — Automatic1111 / Forge / reForge / ComfyUI / NovelAI — accepts it directly when pasted.