Reference to Video

WAN 2.6 Reference-to-Video

Transform video references into new video shots.

模式
Add videos (0/3)

Upload up to 3 videos

Audio: .wav/.mp3, ≤15 MB

Hint: You can drag and drop a file or click to upload

示例

上方显示的是示例输出。提示词和设置已预填在左侧。

您的请求将花费 $1.00 每次运行。 使用 $10 您可以运行此模型约 10 次。

README

Alibaba / WAN 2.6 Reference-to-VideoReference to Video (wan-2.6-reference-to-video)

阿里巴巴 WAN 2.6 参考生成视频将角色、道具或场景参考转换为新的视频镜头,同时保留身份、风格和布局。支持单视角或多视角输入,输出流畅、连贯的运动。

亮点

  • 多视角支持使用 1-3 个参考视频以获得更好的一致性。
  • 身份保留保持角色/对象身份。
  • 流畅运动生成连贯、自然的运动。
  • 高分辨率支持高达 1080P 输出。

参数

  • videos*Reference videos (1-3). Upload video clips showing the character/object/scene.
  • prompt*Describe the scene. Use @Video1, @Video2, @Video3 to reference uploaded videos (e.g. "Dance battle between @Video1 and @Video2").
  • negative_promptThings to avoid in the output video.
  • audioOptional audio track to include in the output.
  • sizeOutput resolution
    • 1280×720 (720P Landscape)
    • 720×1280 (720P Portrait)
    • 1920×1080 (1080P Landscape)
    • 1080×1920 (1080P Portrait)
  • durationVideo duration
    • 5 seconds
    • 10 seconds
  • shot_typeSingle = one continuous shot, Multi = multiple changing shots/cuts
    • Single shot (continuous)
    • Multi-shot (scene cuts)
  • enable_prompt_expansionEnable prompt optimizer for better results
  • seedRandom seed for reproducibility (-1 for random)

定价

$1.00 每次生成

分辨率价格
720P (5s)$1.00
720P (10s)$1.50
1080P (5s)$1.50
1080P (10s)$2.25

如何使用

  1. 1.上传 1-3 个参考视频。
  2. 2.为所需输出编写描述性提示词。
  3. 3.可选择添加负面提示词和音频。
  4. 4.选择分辨率和时长。
  5. 5.生成您的视频。

专业技巧

  • 使用多个视频角度以获得更好的 3D 一致性。
  • 启用提示词扩展以获得更有创意的结果。
  • 使用负面提示词避免不需要的元素。
  • 多视角模式最适合不同的相机角度。

更多可试用的模型

角色视频:创建一致的角色内容。
产品展示:从参考生成产品视频。
场景再现:将参考场景转换为新镜头。
内容扩展:从现有视频生成变体。

Frequently Asked Questions

What is the WAN 2.6 Reference-to-Video API?
Transform video references into new video shots.
How much does WAN 2.6 Reference-to-Video cost via API?
WAN 2.6 Reference-to-Video costs $1.0000 per generation through Renderful's API. No subscription required — pay only for what you use.
How do I use WAN 2.6 Reference-to-Video via API?
Sign up for a free Renderful API key, then send a POST request to the /v1/predictions endpoint with model "wan-2.6-reference-to-video". See the documentation at renderful.ai/docs for code examples in Python, JavaScript, and cURL.
What type of content does WAN 2.6 Reference-to-Video generate?
WAN 2.6 Reference-to-Video is a reference to video model by Alibaba. Key features include: Multi-View Reference, Identity Preservation, Smooth Motion.
Is the WAN 2.6 Reference-to-Video API fast?
WAN 2.6 Reference-to-Video has medium generation speed. Results are delivered via polling or webhook callback for seamless integration.