Toggle Sidebar

Skill PlazaMy Skills

UploadGenerate

All(39)Image(26)Video(4)Audio(2)3D(2)Tools(5)

Unadded only

View detailsAdd skill

y2k-camcorder-look

Apply a fixed "Y2K camcorder" photo filter: an uploaded photo (or a text subject brief) is rendered as a cheap 2003 consumer DV camcorder snapshot — timestamp overlay, blown flash, low-res grain, faint VHS tracking, off-center framing. One image, no choices, no questions. Use when the user uploads a photo and asks for retro / vintage / VHS / camcorder / old-camera / 2000s styling, or sends an "old camcorder photo" subject brief. Opposite register to pastel-y2k-fashion-portrait (candy-pop hyperreal) and Polaroid skills (instant print) — do not bend toward either. Routes through image-executor via image_create (T2I / I2I).

View detailsAdd skill

vintage-patent

Render anything as an authentic late-1800s USPTO patent document — aged ivory paper, fountain-pen marginalia, embossed seal, multi-figure technical drawings. Turns any invention, joke, or uploaded reference image into a "recovered archival filing." Use when the user asks to patent an invention, make a vintage patent-style image or print, convert a photo/drawing into a patent figure, or mock-patent a product/idea as an 1880s filing.

View detailsAdd skill

video-subtitle-extractor

Video subtitle extractor: upload local videos, automatically extract subtitles, and generate transcripts.

View detailsAdd skill

video-generation-skill

Design scripts, storyboards, and editing plans for marketing videos, supporting multi-platform short- and long-form content

View detailsAdd skill

tvc-storyboard

Create a TVC ad storyboard. It's used when users want concepts for commercial films, TVC scripts, storyboard shot lists, 3x3 storyboard grids, visual directions for campaigns, or optional storyboard presentation material based on product briefs, brand briefs, script ideas, or original ad notes.

View detailsAdd skill

text-to-3d-creator

Generate 3D models directly from text. It smartly works with various 3D engines, and you can choose to create reference images first to check the results before turning them into 3D. Great for creating all kinds of 3D assets like characters, props, and scenes.

View detailsAdd skill

style-prompt-engineer

Analyze image styles or receive style descriptions to generate professional style prompts. Supports 27 common styles, automatically adapts between Chinese and English, and the output can be copied and used directly.

View detailsAdd skill

social-poster

Upload product images and selling points to generate marketing posters with professional text layout. Automatically adapts to various platform specifications, with zero changes to product details.

View detailsAdd skill

skill-review

Pre-publish quality review for Skills — structural compliance, description quality, trigger word coverage, best practices audit, scoring + prioritized fixes. 

View detailsAdd skill

sfx-soundscape-creator

Generate environmental sounds, sound effects, ambient sounds, ASMR trigger sounds, game sounds, and film Foley effects based on text descriptions.

View detailsAdd skill

scrapbook-photo-illustration

Turn an uploaded photo into a nostalgic scrapbook-style collage: the original photo kept on top, a tiny hand-painted storybook re-creation of the same scene on warm handmade paper below, plus a small handwritten phrase. Image I2I, fixed vertical 3:4. Use when the user wants their photo turned into this "photo-on-top / mini-illustration-on-paper" memorabilia poster style.

View detailsAdd skill

realistic-landscape

Realistic landscape photography generator. The user describes a scene in natural language (location, camera position, weather, season, elements, etc.); automatically compiles it into one physically coherent, restraint-leaning, low-saturation photo prompt with natural weather (overcast / cloudy / fog / snow / rain preferred) and calls image-executor to produce the photograph. Style: recorded honestly by a camera, not a postcard or saturated promo shot; only lift color when the user explicitly asks for "vivid / bright". Use when the user describes a landscape and wants a photo / realistic landscape photography. Boundary: single landscape photos only (including vertical / multi-image); not characters, products, illustrations, or other image types.

View detailsAdd skill

product-bento-card

Turn product photos into a gorgeous 8-module Bento grid infographic — liquid glass frosted card style, automatically extracting the main product color, perfect for social media promo posts.

View detailsAdd skill

podcast-production-guide

Turn any topic into podcast audio. You can generate a full MP3 with a single click for solo monologues; for multi-person conversations, it outputs separate tracks by role, making it easy to edit later. It can automatically write the script, and after confirmation, generate the audio, in both Chinese and English.

View detailsAdd skill

photo-infographic

Turn photos into beautiful science info graphics with one click. Automatically recognize the subject, extract key facts, and pair them with symmetrical icons and themed color layouts.

View detailsAdd skill

naive-doodle-poster

Make one poster in the naive hand-drawn doodle family — a single big simplified subject, thick wobbly outlines, flat color on a tinted paper field, hand-lettered text. Ships 8 named preset looks (riso print, quote card, sky-blue home life, grounding PSA, rough-marker monster, fantasy scribble, crayon catalog, ink-music). Use when someone wants a charming, imperfect, hand-made-feeling poster, quote card, fake product-catalog page, or mascot announcement — paste a subject and headline, pick a look, with or without a source image. Also 手绘涂鸦海报 / 儿童蜡笔风海报 / 手写语录卡 / 独立杂志 riso 风海报 / 假产品目录页 / 吉祥物公告 / 照片手绘涂鸦化 / 中文字体海报. Not for photoreal, 3D, or clean corporate vector posters.

View detailsAdd skill

megatype-action-poster

Design a big-character action poster — one oversized ultra-condensed headline rendered as a background field that one hero action subject breaks through, as kinetic sports/streetwear/lifestyle campaign key art. Turns a subject, action, prop, and text into a render-ready image_create prompt locked into one original visual system (library of 11 curated looks). Asks which visual system to use when the prompt does not clearly determine one. Trigger when the user wants a sports or streetwear campaign poster, kinetic megatype key art, an oversized-typography action poster, or a big-character poster. Aspect default 16:9; also 9:16, 21:9, 1:1 and other platform-enum ratios.

View detailsAdd skill

jumbotron-broadcast

Create a broadcast-frame "caught on the big screen" image that puts one specific face into a live jumbotron / venue IMAG / crowd-cam frame — realistic because it looks degraded and accidental, not polished. Use when the user wants to look like they were caught on a venue big screen (sports scorebug, concert IMAG, news crowd cut, awards audience cutaway) or makes a feed meme ("me on the jumbotron", "react face on the jumbotron", "did you see me in the audience"). Requires a reference face photo. For a polished studio portrait use editorial-fashion-portrait instead.

View detailsAdd skill

image-prompt-director

Based on reference images, simple descriptions, or a short piece of story, analyze and design a single frame that has a cinematic feel, conceptual expression, or visual appeal.

View detailsAdd skill

ian-xiaohei-illustrations

Turn the judgments, processes, states, and metaphors in a Chinese article into a series of white-background, hand-drawn, quirky but refreshing illustrations for the text.

View detailsAdd skill

headshot-portrait-maker

Generate polished single-subject portraits from uploaded selfies and/or a text prompt — LinkedIn / founder / academic / creator-brand / employee-badge. Output count equals the number of reference selfies (one polished portrait per selfie); with only a text prompt, one portrait is generated. Reference images are optional. Triggers on zh or en: "做个 头像", "生成 LinkedIn 照片", "AI 形象照", "把自拍修成职业照", "职业照", "headshot", "LinkedIn photo", "AI portrait", "professional photo", "founder press shot", "profile pic", or "polish this selfie into a profile photo". Boundary: passport/ID photo, group shot, or swap-background-on-existing-photo belong to other flows — see "When NOT to Use".

View detailsAdd skill

hallmark

"Anti-AI-slop design skill for greenfield pages, audits, redesigns, and design extraction from URLs or screenshots. Use when the user asks to build a new app or landing page, wants to redesign something, invokes Hallmark by name, or uses audit/redesign/study."

View detailsAdd skill

guizang-material-illustration

Create illustrations in the Guisang style: explanatory visuals with annotations, concept diagrams, enhanced charts, data-focused editorial illustrations, educational/scientific mechanism diagrams, humanities/editorial images, and visuals in the style of annotated Chinese materials generated from articles, notes, screenshots, product concepts, reports, teaching materials, or raw data.

View detailsAdd skill

frontend-slides

Create zero-dependency, animation-rich HTML presentations that run purely in the browser. No installation needed—just open the HTML file and you're good to go.

View detailsAdd skill

flat-to-figurine

Turn flat illustrations, line art, and sticker-style images into 3D models. Start by using image-to-image to add volume and lighting; once the user approves the reference, use a 3D engine to generate a high-quality model, completely solving the 'cardboard feel' problem.

View detailsAdd skill

ecommerce-visual-copywriting

Skills in visual planning for e-commerce main images and product detail pages. Start by checking the product info, qualification limits, target platform, and visual references to figure out what drives conversions and lock in the style of the campaign. Then make a storyboard for the main images and detail pages. Once the user approves, provide scene descriptions, on-image text, design notes, and original image prompts.

View detailsAdd skill

ec-visual

Used when users upload product images and visual references to create or improve main product images and product detail pages, while faithfully reproducing the exact product and its visible content provided.

View detailsAdd skill

comic-maker

Turn a story idea into a single multi-panel comic image — manga, American strip, 4-koma, webtoon, or children's-book illustration — generated in ONE image-create call on this platform. Defaults to a 4-panel grid in a single image; the user can request a different panel count in their prompt. Designs the characters, the story beats, the panel order, and any dialogue bubbles all inside one carefully-composed image prompt, then renders the whole comic as a single grid image. Trigger when the user says "make a comic", "draw a comic strip", "generate a manga", "create a 4-panel", "turn this story into a comic", "comic about", or asks for a multi-panel comic image.

View detailsAdd skill

cinematic-video-gen

Generate high-quality, movie-level videos from text prompts or story scripts. It supports storyboard script control, camera movement design, lighting effects, and scene transitions.

View detailsAdd skill

chinese-polisher

Polish Chinese text, correct typos and punctuation, smooth out sentence structures, unify terminology and tone, keep the original meaning and format, with support for light, medium, and heavy edits.

View detailsAdd skill

ccd-2000s-snapshot

Convert an image (i2i) or generate a fresh scene (t2i) in the early-2000s CCD point-and-shoot digital camera look — direct on-camera flash, hard shadows, subdued washed-out colors, lifted flat blacks, blown-out soft highlights, slight cool-neutral white balance drift, mild sensor noise (not film grain), soft edges, subtle chromatic fringing, low dynamic range, nostalgic 2000s everyday snapshot. For i2i it preserves the original subject, pose, and composition. Fixed positive + negative prompts — use them verbatim. Triggers: 2000年代CCD卡片机, CCD卡片机质感, 早期数码相机, CCD相机, 2000s digital camera look, CCD point-and-shoot, 直闪硬影, 传感器噪点.

View detailsAdd skill

brand-moodboard-grid

Create an editorial brand moodboard — a multi-card grid (default 2×2) of distinct editorial cards unified by a named brand's real identity: exact hex colors, typeface DNA, real slogans/products/campaign facts. Use when the user names a brand and wants a brand moodboard, brand-identity grid, lookbook-style brand spread, Hypebeast/Highsnobiety-style brand feature image, or a brand board for a deck.

View detailsAdd skill

brand-kit-video

Turn the brand's logo, products, and colors into a short brand style showcase video—while building or completing the full brand kit.

View detailsAdd skill

brainstorming

Pre-creative brainstorming tool for exploring requirements, proposing solutions, presenting designs, and generating design docs

View detailsAdd skill

baoyu-cover-image

Generate polished cover images for articles with 5-axis customization (type, palette, render, text, mood)

View detailsAdd skill

art-style-transformation-set

Transform one reference photo into a single multi-style portrait grid: the same face rendered in 4-9 distinct art styles (Ghibli, Pixar, oil, 80s anime, LEGO, Polaroid...) in one frame. Locks the reference photo's facial identity and pose; clothing may change per style. Output is ONE grid image (default 6 styles, 3x2; one result only). For multi-style aesthetic renders, "every version of me" posts, gifts, personal-brand visuals. Reference photo preferred (I2I) for identity lock; falls back to a one-sentence description (T2I).

View detailsAdd skill

app-icon-concept-gen

Generate a set of logo designs. The default layout is a 2x2 grid featuring four logo variations; the grid layout adjusts automatically based on the requested quantity. Upon rendering, explain the design concept behind each logo and include a note regarding trademark usage. Selection criteria: Designs should be flat, minimalist, visually focused, and easily recognizable, without incorporating specific text. Trigger commands: "Generate logos," "Create a few logos," "Brand logo," "Logo concepts," "Create a logo based on this website (URL)," "Logo set." Functional scope: This skill generates artistic logo concepts; authorization is required for the reuse of protected trademarks; this skill is not intended for creating UI icon sets or CMYK print output.

View detailsAdd skill

ai-travel-snapshot

Put a real face photo into a famous landmark and produce an "I'm on vacation here" AI travel snapshot. Keeps the same person across a batch of 2–4 destinations (#aivacation meme, must be labeled as AI travel). Triggers: "put me in Paris", "send me to XX", "AI travel photo", "fake my vacation", a cute travel pic for a friend, birthday/anniversary "greetings from {a place they'd love}". Boundary: if the user wants a real itinerary/planning guide, use the travel-planner skill (or make that skill). This skill only produces hero visual images.

View detailsAdd skill

90s-film-snapshot

Convert an image (i2i) or generate a fresh scene (t2i) in the 1990s Japanese analog film snapshot look — shot on 35mm Fujifilm Superia 400 / Kodak Gold negative. For i2i it preserves the original subject, pose, composition, and framing, changing only the visual style. Fixed positive + negative prompts — use them verbatim, never paraphrase. Triggers: 90s日系胶片, 90年代胶片风, 日系胶片快照, Fujifilm Superia, Kodak Gold, analog film snapshot, 90s film look.