Turn everyday photos into art, with AI.

_{Pick a style, send a photo, get a finished artwork. No prompt engineering, no parameter tuning.}

36 individually installable visual style Skills · Try online · Community · License

Created and maintained by @威比 Hunter Wei. (same handle on Douyin & Xiaohongshu)
English · 简体中文

---

For AI Agents: This repository provides modular, production-ready visual style Skills for OpenAI Codex. The user selects a style, supplies a photo, and receives a completed design artifact.

• Available Skills: See the All styles directory below, organized by subject with style names, invocation slugs, and photo recommendations.
• How to install: Provide https://github.com/Vieeeeeee/wibi-style/tree/main/skills/<call-name> and have the user run or install that URL in Codex. Each style resides in an isolated directory; only the selected Skill is downloaded.
• How to use: In a new task, submit Use $call-name on this photo. The Skill introduces itself with an interactive welcome card on its first turn.

Style overview

Click any thumbnail to jump to that Skill's dedicated directory. You can also preview them interactively at Wibi Style Lab.

Electric Blue Halftone Poster	Bold Line Manga Avatar	Crayon Sketch Avatar	Dark Red & Black Cel Shade	Iridescent Long Exposure
Blue Retro Print	Underground Audition Skill	Diamond Grin Kid Portrait	Childhood Head Card	Fisheye City Poster
Niu Lai Movie Poster	Marker-Pen Child Doodle	Quirky Pop Doodle Sticker	Photo Perler Charm	Pixel Slice Stretch
Glitch Pixel Collage	Clear Sky Urban Cel	Retro Table Magazine	Framed Comic Panel	Textile Toy Portrait
Cold Blue Glitch Anime	Crayon Flat Naive	Office Animals	Vintage Home Journal	Serpent Line Ink
Warm Paper Crayon Street	Minimal Pet Line Art	Museum Impression Oil	Urban Mecha Cel	Primary Hardlight City
Personalized Print Ticket	Paper Flat Vignette	Lo-Fi Print Anime	复古扁平动画头	Red & Blue Fashion Manga
Cream Paper Doodle

All styles

36 styles right now — find yours by the photo you have. Click a name for its detail page; the call name is both the last part of the install URL and the $ command you use.

Subject	Your photo	Styles (click the name for details · code is the call name)
Portraits & selfies	Selfies, avatars, close-up people shots	Electric Blue Halftone Poster electric-blue-halftone-poster Single-person portraits or pet close-ups Bold Line Manga Avatar alt-manga-avatar Front-facing or three-quarter selfies Crayon Sketch Avatar art-print-poster Selfies with clear features and a memorable expression Dark Red & Black Cel Shade dark-red-black-cel-shaded Portraits that suit dramatic red-and-black lighting Iridescent Long Exposure iridescent-long-exposure Close-ups where you want a hazy, dreamlike mood Blue Retro Print blue-retro-print Portraits with clear outlines and expression Underground Audition Skill underground-audition Single-person half-body or close-up photos for a cool direct-flash turn-card sequence Textile Toy Portrait textile-toy-portrait Clear solo portraits with readable hair, expression, accessories, or hand interaction Cold Blue Glitch Anime cold-blue-glitch-anime Portraits, selfies, street snapshots, or nightlife photos with expressive poses Crayon Flat Naive crayon-flat-naive Portraits, half-body, full-body, or seated photos where clothing and pose are clear Serpent Line Ink (beta) serpent-line-ink Single-person head-and-shoulders, close-up, or half-body portraits — selfies or mirror selfies both work Warm Paper Crayon Street warm-paper-crayon-street Street snaps, full-body or half-body portraits; half-body shots keep their crop 复古扁平动画头 retro-flat-anime-head Clear solo portraits, front or three-quarter view, with readable hair and expression Cream Paper Doodle cream-paper-doodle Expressive portraits, everyday gestures, and people with pets on warm ivory paper
Portraits · lo-fi avatars	Close-up solo portraits and selfies	Lo-Fi Print Anime lofi-print-anime Close-ups with readable expression, hair, glasses, or hats
Kids & childhood	A child's face, or your own childhood look	Diamond Grin Kid Portrait diamond-kid-head-card Single-child photos where face, hair, or hat is recognizable Childhood Head Card kid-head-card Old childhood photos; modern selfies are aged back to 5–8
Group shots & posters	Group photos, street shots, travel photos	Fisheye City Poster (beta) fisheye-city-cover Into a Y2K heavy-fisheye city cover Niu Lai Movie Poster niulai-movie-poster Landscape solo, duo or trio photos into a low-budget old-animation cow-head Chinese film poster
Pets	One cat or dog with a clear pose, gaze, and body outline	Minimal Pet Line Art minimal-pet-doodle Turns a memorable gesture into a few black lines, one fur-color smudge, and expansive paper space
Anything goes	People, pets, products, objects — all fine	Marker-Pen Child Doodle marker-child-doodle Real photos with clear people, pets, and props; half-body crops stay half-body Quirky Pop Doodle Sticker quirky-pop-doodle-sticker Clear subject with a memorable pose — people, pets, products, props Photo Perler Charm photo-perler-charm People, pets, bouquets, food, and objects with clean silhouettes Pixel Slice Stretch pixel-stretch People, still life, or anything with a clear subject Glitch Pixel Collage glitch-pixel-collage People, still life, or photos with distinct color layers Office Animals office-animals Specify an animal and office scene to generate a Y2K absurd work poster Personalized Print Ticket personalized-print-ticket People, travel, streets, architecture, food, and everyday photos with custom bilingual copy in one of four print-ticket styles
Scenes & food	Streets, architecture, the dinner table, countryside, and coastlines	Clear Sky Urban Cel clear-sky-urban-cel City streets, architecture, travel documentary, transit Primary Hardlight City primary-hardlight-city City streets, architecture, public transit, travel documentary, and people in urban environments Retro Table Magazine retro-table-print Table photos where dishes and vessels are readable Museum Impression Oil museum-impression-oil Landscape photos of streets, villages, roads, hills, riverbanks, and coastlines Paper Flat Vignette paper-flat-vignette Street snaps, cafe, bookstore, travel, and everyday portraits composed as centered flat-paper vignettes
Home & interiors	Corners of your home — living room, bedroom, shelves, coffee bar	Vintage Home Journal home-journal Any casual corner of your home — living room, bedroom, coffee bar, bookshelf
Advanced · multi-step	When you want to crop the best part first	Framed Comic Panel wibi-frame Photos with clear eye contact, expression, gesture, or object relationships — crop first, then generate
Portraits · expressive moments	Single-person photos with a memorable expression or gesture	Urban Mecha Cel urban-mecha-cel Clear selfies, casual snapshots, and outtakes turned into rough red-on-white manga drawings Red & Blue Fashion Manga red-blue-fashion-manga Solo outfit shots and street portraits; square manga poster with photo-inspired lettering

Wibi Style · is a display prefix only. English Skill names, GitHub URLs, and $call-names never carry it.

Install and use

1. Send the install command in Codex

Swap the call name in the URL for the style you picked (Electric Blue Halftone Poster shown here):

Please install this Skill:
https://github.com/Vieeeeeee/wibi-style/tree/main/skills/electric-blue-halftone-poster

The installer downloads only that one Skill. Nothing else to do — the first time you call it, it introduces itself and tells you what's next.

2. Start a new task, upload a photo, and call it

Use $electric-blue-halftone-poster on this photo

For any other style, swap $electric-blue-halftone-poster for its call name from the table above.

What's included and how updates work

Every style ships its own SKILL.md, runtime rules, prompt, version manifest, and whatever reference material or support files that style needs. Nothing is shared or mixed between styles.

The first time you call a style in a new task, it checks whether that one style has a new version — nothing else. It only tells you when a newer version exists; it never overwrites your local files, and it uploads no photos and no usage data.

Community

威比 😌 AIGC 学习群

A group for AI visual techniques, Skill usage questions, and new style previews

The group has reached 200 members, so direct QR code scanning is no longer available. Please add WeChat Wibi2077 with the note 「进群」 to be invited.

Heads up: this is a WeChat group and the conversation is in Chinese. It needs the WeChat app, so it may not be practical outside mainland China. If you can't join, open an issue instead — that works from anywhere.

Skills that support the group entry read the root community.json on demand. This lookup uploads no photos and no usage data.

License and attribution

When you share generated results publicly, please credit:

Visual Skill by @威比 Hunter Wei.

This repository follows Component-Specific Licensing (see LICENSE):

• Code and Tooling (MIT License): All automation scripts, test suites, manifest schemas, and CI workflows are freely available under the permissive MIT License.
• Original Visual Skills & Prompts (Personal Non-Commercial License): Original Skill logic, prompts, styling directives, and composition guides are licensed for personal, non-commercial use only. Commercial use, resale, paid redistribution, or commercial SaaS integration requires prior written permission from the author.

Each Skill's SOURCES.md states exactly which rights the package does and does not hold. Third-party reference assets are bundled solely for visual-style guidance; rights remain with their respective rights holders.