Back to Details

muapi-fashion-try-on

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Fashion Try-On

时尚虚拟试穿

Virtually try on different outfits by combining a person's photo and a clothing item, then optionally generate a professional fashion model video.
Estimated credits: ~150 per run.
通过结合人物照片与服装单品实现虚拟试穿不同穿搭，还可选择生成专业时尚模特视频。
预估消耗积分： 每次运行约150积分。

Inputs

输入项

NameTypeRequiredDefaultDescription
person_image
image_urlyes—A photo of the person or model who will try on the clothes.
clothing_image
image_urlyes—A photo of the clothing item to try on.
名称类型是否必填默认值描述
person_image
image_url是—将要试穿服装的人物或模特照片。
clothing_image
image_url是—待试穿的服装单品照片。

Steps

步骤

Phase A — Virtual Try-On

阶段A——虚拟试穿

If 
{{person_image}}
or 
{{clothing_image}}
is not provided, ask the user to upload them.
Once both images are available, submit the plan with ONE step to perform the try-on:
  1. Fashion Try-On — 
    muapi image edit
    (model=
    qwen-image-edit-2511
    ):
    • Reference Images: Use both 
      {{person_image}}
      and 
      {{clothing_image}}
      .
    • Prompt: 
      A high-quality fashion photograph of the person from the first reference image wearing the exact clothing item from the second reference image. The fit should be natural and realistic, maintaining the person's pose and the clothing's texture and patterns. Soft studio lighting, neutral background, professional fashion photography style.
    • Aspect ratio: 1:1 or 4:5
Present the resulting fashion photo to the user for approval.
若未提供
{{person_image}}
或
{{clothing_image}}
，请要求用户上传相应图片。
获取两张图片后，提交包含以下单个步骤的试穿计划：
  1. 时尚虚拟试穿 —— 
    muapi image edit
    （模型=
    qwen-image-edit-2511
    ）：
    • 参考图片：同时使用
      {{person_image}}
      和
      {{clothing_image}}
      。
    • 提示词：
      一张高质量时尚照片，展示第一张参考图片中的人物穿着第二张参考图片中的同款服装。服装贴合度自然逼真，保留人物姿势以及服装的纹理和图案。采用柔和的影棚灯光、中性背景，风格为专业时尚摄影。
    • 宽高比：1:1或4:5
将生成的时尚照片展示给用户确认。

Phase B — Fashion Video Generation (Optional)

阶段B——时尚视频生成（可选）

After the image is generated, ask the user if they would like to create a professional fashion video of the model wearing the outfit.
If requested, submit the plan with ONE step:
  1. Fashion Video Generation — 
    muapi video from-image
    (model=
    seedance-v1.5-pro-i2v-fast
    ):
    • Reference Image: The try-on image generated in Phase A.
    • Prompt: 
      Shot type of Three-Quarter Length Shot. [Push in] as model gracefully places hand on hip, shifts weight to one side, tilts head slightly with soft smile, and gently adjusts hair with fingertips, creating elegant movement and confidence.
    • Aspect ratio: 9:16 or 4:5
After generation, present the final fashion video.
生成试穿图片后，询问用户是否需要创建模特穿着该穿搭的专业时尚视频。
若用户同意，提交包含以下单个步骤的计划：
  1. 时尚视频生成 —— 
    muapi video from-image
    （模型=
    seedance-v1.5-pro-i2v-fast
    ）：
    • 参考图片：阶段A生成的试穿图片。
    • 提示词：
      镜头类型为四分之三身长镜头。模特优雅地将手放在臀部，将重心移至一侧，微微歪头并露出柔和微笑，用指尖轻轻整理头发，呈现出优雅的动作与自信姿态，镜头[推进]。
    • 宽高比：9:16或4:5
生成完成后，展示最终的时尚视频。

Trigger Keywords

触发关键词

fashion try on
, 
virtual fitting room
, 
try on clothes
, 
model fashion
, 
clothing preview
fashion try on
, 
virtual fitting room
, 
try on clothes
, 
model fashion
, 
clothing preview

Notes for the Executing Agent

执行Agent注意事项

  • This recipe is LLM-orchestrated: read each phase, gather any missing inputs from the user, then call 
    muapi
    CLI commands. Use 
    muapi auth configure
    first if 
    MUAPI_API_KEY
    is unset.
  • For model IDs without a CLI alias yet, fall back to the raw endpoint via 
    curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'
    and poll with 
    muapi predict wait <request_id>
    .
  • Substitute 
    {{input_name}}
    placeholders with the user's actual inputs before issuing each call.
  • 本流程由LLM编排：阅读每个阶段，向用户收集缺失的输入项，然后调用
    muapi
    CLI命令。若
    MUAPI_API_KEY
    未设置，请先执行
    muapi auth configure
    。
  • 对于尚未设置CLI别名的模型ID，可通过原始端点调用：
    curl -X POST https://api.muapi.ai/api/v1/<endpoint> -H "x-api-key: $MUAPI_API_KEY" -H 'content-type: application/json' -d '{...}'
    ，并使用
    muapi predict wait <request_id>
    轮询结果。
  • 在发起每次调用前，将
    {{input_name}}
    占位符替换为用户的实际输入内容。