yang0/handraw-style

handraw‑style – A Prompt‑Library for Hand‑Drawn Image Styles

What it is

• A curated collection of 216 hand‑drawn illustration styles, each assigned a short numeric ID (001‑216).
• For each style the repository provides a short name in Chinese and English and a ready‑made prompt that can be fed directly to text‑to‑image models (e.g., Stable Diffusion, Midjourney, etc.).
• A tiny web UI (gallery/index.html) lets you browse thumbnail samples and copy the corresponding ID.

Who it helps

• Creators who struggle to describe a visual style in words (social‑media posts, short‑form video, branding).
• Anyone who wants consistent visual “tone” across many generated images without memorising style names.

How it works

1. Open the gallery page and pick a style you like; note its number (e.g., 041).
2. Combine the number with your subject, e.g. 041号风格，主题：秋天的第一杯奶茶.
3. The repository’s Skill (a simple script or chatbot integration) returns a bilingual prompt that includes:
   • The style’s official name (Chinese + English)
   • Core descriptive keywords for that style
   • Optionally a reference image (the thumbnail) if the model cannot be triggered by the name alone.
4. Paste the prompt into any “text‑to‑image” service that supports image generation.

Why it matters

• Stability – Using a fixed ID removes the “I said cute, you gave me gritty” problem caused by vague adjectives.
• Speed – No need to craft long descriptive prompts; the library does it for you.
• Cross‑model consistency – The system first tries to activate the style by name; if that fails it supplies the reference thumbnail, preventing the model from locking onto unwanted composition details.

Structure of the repo

• gallery/ – HTML page with a searchable grid of style thumbnails.
• images/ – PNG sheets that show groups of styles (A‑G categories). Each sheet contains the visual examples for a range of IDs.
• Prompt data (style name, English translation, keyword list) is embedded in the Skill’s code (not shown in the README) and is indexed by the numeric IDs.

Typical usage examples

041号风格，主题：秋天的第一杯奶茶   →  "Hand‑drawn cute style, warm pastel tones, soft lines…"
210号风格，主题：小男孩在雪地里点鞭炮 →  "Japanese illustration, dynamic brushwork, bright fireworks…"
193号风格，主题：大唐夜宴           →  "Chinese historical illustration, rich colors, intricate costumes…"

Bottom line yang0/handraw-style is a practical, AI‑adjacent tool that bridges the gap between human artistic intent and the often‑opaque prompt language required by modern generative‑image models. It’s a lightweight, non‑code‑heavy resource that can be dropped into any workflow that uses text‑to‑image generation.