Anime Outfit to Real Woman · skill Open Skill công khai
Anime Outfit to Real Woman
Anime Outfit to Real Woman là một Open Skill công khai trên Digen. Mở ra để xem workflow này tạo gì.
Doc
0 0 1
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4150474/2026-10-15/d8ypaae4.jpg
# Anime Wardrobe to Real Woman Turns one complete anime/illustrated outfit into a single 3:4 poster: the outfit extracted as flat anime garment illustrations **plus** the same wardrobe photographed on a photorealistic adult woman. One tool call, one image out. No structured project state — everything the run needs (garment list, palette, wearer identity) is decided up front and written into the generation prompt. ## When to use - User describes an outfit in text and wants to see it on a real woman. - User uploads an anime character / anime outfit / flat-lay garment reference. - User uploads a photo of a woman and wants her to wear the anime wardrobe. - User asks to redo a previous result with a different color, garment, or wearer. ## Inputs | Input | Role | Notes | |---|---|---| | Text | Outfit concept, colors, materials, mood | Always accepted, even when images are attached | | Reference image: **Anime Outfit** | Controls the **wardrobe** (garment design, silhouette, colors) | Anime character wearing it, or a flat-lay of the garments | | Reference image: **Adult Woman** | Controls the **wearer identity** (face, hair, skin tone, build) | A plain portrait photo | Keep the two reference roles separate: wardrobe comes from the outfit image, identity comes from the woman image. When only the outfit image is given, invent a natural adult wearer (25–40) whose hair and coloring suit the palette. When only the woman image is given, invent an ordinary everyday outfit in a unified warm palette. ## Output format (fixed) Vertical **3:4** poster on a warm off-white studio background, five zones: 1. **Top** — two-line serif headline `Anime Wardrobe` (deep teal) / `to Real Woman` (brick red) with a small wheat-leaf divider. 2. **Left third** — full-body **anime illustration** of an adult woman wearing the whole outfit. 3. **Center column** — rounded cream panel titled `OUTFIT ITEMS`, every garment drawn separately as a clean flat-lay illustration, stacked vertically, each with a small uppercase serif caption (`CARDIGAN`, `CREWNECK TOP`, `PLEATED SKIRT`, `SOCKS`, `LOAFERS`, …). 4. **Right third** — **photorealistic photo** of an adult woman wearing exactly the same outfit. 5. **Bottom left** — tan-bordered box titled `ORIGINAL INPUT` with two rounded thumbnails labelled `1. Anime Outfit` (flat-lay of the garments) and `2. Adult Woman` (headshot of the same woman); rust-orange curved arrows lead from the box to the center panel and to the photoreal woman. Garment design, color, and material must read as the **same object** in all three places: anime figure, flat-lay items, and photo. ## Steps 1. **Lock the spec.** List every garment (type, color name, material, silhouette), the palette (2–4 colors), and the wearer (adult 25–40, hair color/length, skin tone, build). If the user gave a garment list, keep their wording; otherwise derive it from the reference image. 2. **Read the reference images.** Use what is visible in the attached images directly. If a garment's cut, closure, or color is not readable, call `recognize_image(image_url=<reference>, prompt="Describe each garment: type, cut, closure, color, fabric, and any print, plus the wearer's hair color, length, and skin tone.")` before writing the prompt. Never guess details you could not see and then present them as the user's design. 3. **Compose the generation prompt in English** using the template below, filling every slot from step 1–2. Do not shorten it to a one-line idea — the layout, both style blocks, and the garment labels are what make the poster readable. 4. **Generate.** Pass the reference images in `input_urls` in the order [Anime Outfit, Adult Woman] (send only the ones supplied). 5. **Reply** in one or two natural sentences (which outfit, which wearer, any choice you invented). Do not paste the image URL. 6. When the poster is delivered, optionally call `set_guidance` in the same turn with 2–3 concrete follow-ups (e.g. `Change the cardigan to navy`, `Put this outf