Clothes Try-On Studio

Virtual clothes try-on studio using YouCam (Perfect Corp) AI. Swap outfits onto the user's photo, optionally change the background, and optionally turn the result into a short motion video (turn / runway / pose). Use for "換衣", "虛擬試穿", "outfit try-on", "try on clothes". Do NOT use for makeup, hair, or skin analysis.

A generation skill: try on clothes, optionally change the background, optionally turn the result into a short motion video. There is no run.py.

When to trigger

When the user wants to try clothes on virtually. For makeup, hair, or skin, hand off to the matching skill.

Ask first (this skill has multiple generation steps) and state cost

Before generating, ask the user which steps they want and tell them the credit cost of each:

• Try on clothes (cloth-v4) — required base step
• Change background (bg-replace) — optional
• Motion video (image-to-video) — optional; note video is billed per second Get costs with python scripts/youcam_core.py cost --feature cloth-v4 (and --feature bg-replace, --feature image-to-video), tell the user the total, and only run what they choose.

Get the latest API spec (do this every run)

• Clothes: https://docs.perfectcorp.com/reference/ai_clothes.md (latest engine cloth-v4)
• Background change: https://docs.perfectcorp.com/reference/ai_photo_background_change.md
• Video generator: https://docs.perfectcorp.com/reference/ai_video_generator.md Pass --version if newer than api-fallback.yaml.

Prerequisites

1. A photo of the person (full-body or upper-body) plus one garment reference image (product or worn photo) or a template_id.
2. python scripts/youcam_core.py validate-key (0 credits); guide the user to set the key if missing.

Flow (only the chosen steps)

1. Try on (cloth-v4):

   text

   python scripts/youcam_core.py run --feature cloth-v4 --src_file <person.jpg> --ref_file <garment.jpg> \
       --param garment_category=full_body
   

   (or use a template: --param template_id=<id>; confirm the exact garment_category value from the latest doc — see Notes.)
2. (Optional) Change background (bg-replace): on the result or original image, run --feature bg-replace --src_file <img> --param type=prompt --param prompt="A sunny beach" or --param type=template --param template_id=<id>.
3. (Optional) Motion video (image-to-video): run --feature "image-to-video" --src_file <img> --param resolution=720 --param dst_duration=5 --param prompt="turn around / runway walk"
4. Each generation returns a result image/video URL in data.results.url.

Output rules

• Return the result image/video URL exactly as given (hyperlink is OK; do not modify or shorten it).
• Video takes longer (the poll timeout is set higher); wait for success.
• Always end with this note: "This try-on image/video is AI-generated and for reference only; actual fit and appearance may vary."

Notes

cloth-v4 garment_category is confirmed for full_body; the upper/lower-body token spelling is not yet 100% confirmed, so verify it against the latest doc parameter table before running (see the handoff GAPS).