创建 Wan 视频 - CometAPI Documentation

POST

videos

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F model=wan3.0 \
  -F 'prompt=A small robot walks through a bright greenhouse with smooth camera movement.' \
  -F seconds=5 \
  -F size=1920x1080

{
  "id": "task_example",
  "task_id": "task_example",
  "object": "video",
  "model": "wan3.0",
  "status": "queued",
  "progress": 0,
  "created_at": 1779938152
}

使用文本、参考媒体或首尾帧创建 Wan 视频。保存返回的 id，以便轮询任务并下载已完成的视频。 POST /v1/videos 使用 multipart/form-data。请以 seconds 的形式发送明确的整数值，并以 size 的 WxH 格式发送。

选择输入模式

每种模式都需要 model 和 prompt。可用的媒体输入取决于模型：

目标	模型 ID	媒体字段
文本生成视频	`wan2.6`, `wan2.7`, `wan3.0`, `wan3.0-prime`	省略媒体字段
图像生成视频	`wan2.7`, `wan3.0`, `wan3.0-prime`	`input_reference` URL 或上传的图像文件
参考视频	`wan2.6`	`reference_videos` 仅
图像和视频参考	`wan2.7`, `wan3.0`, `wan3.0-prime`	`input_reference` 搭配 `reference_videos`
图像、视频和音频参考	`wan3.0`, `wan3.0-prime`	图像/视频参考，搭配 `reference_audios`
首尾帧	`wan2.7`, `wan3.0`, `wan3.0-prime`	`first_frame`，可选 `last_frame`

要查找可供您的 API 密钥使用的 Wan 模型 ID，请使用列出可用模型.

使用参考图像

对于 wan2.7、wan3.0 和 wan3.0-prime，请通过 input_reference 以可公开访问的 HTTPS URL 或上传文件的形式发送参考图像。对于多张图像，请按图像顺序重复该字段，每个字段提供一个 URL 或文件。在 input_reference 中使用全部 URL 或全部文件。请分别发送每个 URL，而不要将多个 URL 合并为逗号分隔的值。使用每张最大 20 MB 的 PNG 或 JPEG 图像。单个 input_reference URL 或文件将使用图像生成视频工作流。该图像会引导构图和外观；请使用 Prompt 描述运动和摄像机行为。对于 wan3.0 和 wan3.0-prime，无论使用 URL 还是上传文件，均最多通过 input_reference 发送 10 张参考图像。

使用参考视频和音频

通过 reference_videos 发送视频，并通过 reference_audios 发送音频片段；每个输入均需重复传入相应字段一次。每个输入都可以是可公开访问的 HTTPS URL 或上传的文件。视频参考请使用包含 H.264 编码视频的 MP4 文件。音频参考请使用单独的 MP3 文件。不同字段可以使用不同的输入类型，例如将图片 URL 与上传的视频搭配使用。在 reference_videos 或 reference_audios 中，请全部使用 URL 或全部使用文件。最多可上传五个视频文件，每个最大为 50 MiB（50 × 1024 × 1024 字节). 最多可上传五个音频文件，每个最大为 15 MiB（15 × 1024 × 1024 字节).

Wan 2.6 视频参考

发送一到三个 reference_videos，每个时长为 2–30 秒，大小不超过 30 MB，且帧率至少为 16 fps。这些限制同时适用于 URL 和上传的视频。在 Prompt 中，按照输入顺序将主体称为 @Video1、@Video2 和 @Video3。在此模式下，省略 input_reference、reference_audios、first_frame 和 last_frame。将 seconds 设置为 5 或 10。以下示例根据两个视频参考创建一个五秒视频。请将媒体 URL 或文件名替换为你的输入：

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.6' \
  --form-string 'prompt=Keep @Video1 and @Video2 recognizable as they walk.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'reference_videos=https://your-media-host/reference-1.mp4' \
  -F 'reference_videos=https://your-media-host/reference-2.mp4'

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.6' \
  --form-string 'prompt=Keep @Video1 and @Video2 recognizable as they walk.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'reference_videos=@reference-1.mp4;type=video/mp4' \
  -F 'reference_videos=@reference-2.mp4;type=video/mp4'

Wan 2.7 图片和视频参考

结合图片和视频参考来引导主体外观和动作。将 seconds 设置为从 2 到 10 的整数。省略 reference_audios。使用两个图片 URL 或两张上传的图片，并搭配一个视频参考：

curl \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.7' \
  -F 'prompt=Preserve the image subjects and follow the video motion.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'input_reference=https://your-image-host/reference-1.png' \
  -F 'input_reference=https://your-image-host/reference-2.jpg' \
  -F 'reference_videos=https://your-media-host/reference.mp4'

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.7' \
  -F 'prompt=Preserve the image subjects and follow the video motion.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'input_reference=@reference-1.png;type=image/png' \
  -F 'input_reference=@reference-2.jpg;type=image/jpeg' \
  -F 'reference_videos=@reference.mp4;type=video/mp4'

Wan 3.0 图片、视频和音频参考

对于 wan3.0 和 wan3.0-prime，结合图片和视频参考以及音频参考，以引导外观、动作和节奏。最多发送五个视频参考，总时长不超过 15 秒。每个视频的帧率必须至少为 16 fps。最多发送五个音频参考，总时长不超过 15 秒。视频和音频时长分别计算。将输出 seconds 设置为从 2 到 30 的整数。使用两个图片 URL 或两张上传的图片，并搭配一个视频和两个音频片段：

curl \
  https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan3.0' \
  -F 'prompt=Preserve image subjects, video motion, and audio rhythm.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'input_reference=https://your-image-host/reference-1.png' \
  -F 'input_reference=https://your-image-host/reference-2.jpg' \
  -F 'reference_videos=https://your-media-host/reference.mp4' \
  -F 'reference_audios=https://your-media-host/reference-1.mp3' \
  -F 'reference_audios=https://your-media-host/reference-2.mp3'

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan3.0' \
  -F 'prompt=Preserve image subjects, video motion, and audio rhythm.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'input_reference=@reference-1.png;type=image/png' \
  -F 'input_reference=@reference-2.jpg;type=image/jpeg' \
  -F 'reference_videos=@reference.mp4;type=video/mp4' \
  -F 'reference_audios=@reference-1.mp3;type=audio/mpeg' \
  -F 'reference_audios=@reference-2.mp3;type=audio/mpeg'

使用首帧和尾帧

对于 wan2.7、wan3.0 和 wan3.0-prime，为起始图像发送 first_frame，并可选地为结束图像发送 last_frame。每个字段都接受一个可公开访问的 HTTPS PNG 或 JPEG URL，或者一个已上传的 PNG 或 JPEG 文件。对于 Wan 2.7 帧输入，起始图像决定宽高比，而 size 选择输出分辨率级别。请将帧输入与 input_reference、reference_videos 和 reference_audios 分开提交请求。以下 Wan 2.7 示例会在起始帧和结束帧之间创建过渡：

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.7' \
  -F 'prompt=Move smoothly from the opening frame to the ending frame.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'first_frame=https://your-image-host/first-frame.png' \
  -F 'last_frame=https://your-image-host/last-frame.jpg'

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan2.7' \
  -F 'prompt=Move smoothly from the opening frame to the ending frame.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'first_frame=@first-frame.png;type=image/png' \
  -F 'last_frame=@last-frame.jpg;type=image/jpeg'

以下 Wan 3.0 示例会在起始帧和结束帧之间创建过渡：

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan3.0' \
  -F 'prompt=Move smoothly from the opening frame to the ending frame.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'first_frame=https://your-image-host/first-frame.png' \
  -F 'last_frame=https://your-image-host/last-frame.jpg'

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F 'model=wan3.0' \
  -F 'prompt=Move smoothly from the opening frame to the ending frame.' \
  -F 'seconds=5' \
  -F 'size=1280x720' \
  -F 'first_frame=@first-frame.png;type=image/png' \
  -F 'last_frame=@last-frame.jpg;type=image/jpeg'

设置时长和尺寸

将 seconds 作为整数发送，其值须在适用于你的模型和输入模式的范围内。

模型 ID	输入模式	`seconds`	默认值
`wan3.0`, `wan3.0-prime`	文本、参考素材或帧	整数 `2`–`30`	`5` 秒， `1920x1080`
`wan2.7`	文本	整数 `2`–`15`	`5` 秒， `1920x1080`
`wan2.7`	帧	整数 `2`–`15`	`5` 秒，`1080p` 层级；宽高比遵循起始图像
`wan2.7`	图像/视频参考素材	整数 `2`–`10`	`5` 秒， `1920x1080`
`wan2.6`	文本	`5`, `10`, `15`	`5` 秒， `1920x1080`
`wan2.6`	视频参考素材	`5`, `10`	`5` 秒， `1920x1080`

将 size 设置为适用于目标模型的以下 WxH 值之一。 Wan 3.0 的尺寸值也适用于 wan3.0-prime。

Wan 3.0 尺寸值

分辨率层级	宽高比	`size` (`WxH`)
`480p`	`16:9`	`832x480`
	`9:16`	`480x832`
	`1:1`	`624x624`
`720p`	`16:9`	`1280x720`
	`9:16`	`720x1280`
	`1:1`	`960x960`
	`4:3`	`1088x832`
	`3:4`	`832x1088`
`1080p`	`16:9`	`1920x1080`
	`9:16`	`1080x1920`
	`1:1`	`1440x1440`
	`4:3`	`1632x1248`
	`3:4`	`1248x1632`

Wan 2.7 尺寸值

分辨率层级	宽高比	`size` (`WxH`)
`720p`	`16:9`	`1280x720`
	`9:16`	`720x1280`
`1080p`	`16:9`	`1920x1080`
	`9:16`	`1080x1920`
	`1:1`	`1440x1440`

Wan 2.6 尺寸值

分辨率层级	宽高比	`size` (`WxH`)
`720p`	`16:9`	`1280x720`
	`9:16`	`720x1280`
	`4:3`	`1088x832`
	`3:4`	`832x1088`
`1080p`	`16:9`	`1920x1080`
	`9:16`	`1080x1920`
	`1:1`	`1440x1440`
	`4:3`	`1632x1248`
	`3:4`	`1248x1632`

任务流程

创建任务

发送 multipart 表单请求，并存储返回的 id。

轮询任务

调用检索 Wan 视频直到 status 为 completed、failed 或 error。

下载结果

当任务状态为 completed 时，调用检索 Wan 视频内容以下载 MP4 文件。

授权

Authorization

string

header

必填

Bearer authentication. Use your CometAPI API key.

请求体

multipart/form-data

model

string

必填

Wan model ID: wan2.6, wan2.7, wan3.0, or wan3.0-prime. Use wan2.6 for text or reference-video input. Use wan2.7 for text, image/video references, or frame inputs. Use wan3.0 or wan3.0-prime to also include reference audio.

示例:

"wan3.0"

prompt

string

必填

Text prompt that describes the video to generate. Describe the appearance and motion to preserve from references, or the transition between first_frame and last_frame. For wan2.6 video references, use @Video1, @Video2, and @Video3 in input order.

示例:

"A small robot walks through a bright greenhouse with smooth camera movement."

seconds

integer

Requested output duration in whole seconds. Send an explicit integer. For wan2.6 reference-to-video, use 5 or 10; text-to-video also accepts 15. For wan2.7 reference-to-video, use 2 through 10; text and frame inputs accept 2 through 15. For wan3.0 and wan3.0-prime, use 2 through 30. Default is 5.

必填范围: 2 <= x <= 30

示例:

5

size

string

Supported WxH size values. For wan3.0 and wan3.0-prime: 832x480, 480x832, 624x624, 640x480, 480x640, 1280x720, 720x1280, 960x960, 1088x832, 832x1088, 1920x1080, 1080x1920, 1440x1440, 1632x1248, 1248x1632. For wan2.6: 1280x720, 720x1280, 1088x832, 832x1088, 1920x1080, 1080x1920, 1440x1440, 1632x1248, 1248x1632. For wan2.7: 1280x720, 720x1280, 1920x1080, 1080x1920, 1440x1440. Default is 1920x1080. For Wan 2.7 frame inputs, the opening image determines the aspect ratio; size selects the output resolution tier.

示例:

"1920x1080"

input_reference

PNG or JPEG reference images for wan2.7, wan3.0, and wan3.0-prime. Repeat input_reference with one publicly accessible HTTPS image URL or uploaded image file per field. Use all URLs or all files within this field. Do not combine URLs into a comma-separated value. A single URL or file uses the image-to-video workflow. Each image can be up to 20 MB. For wan3.0 and wan3.0-prime, send at most 10 reference images, whether you use URLs or uploaded files. Image and video references can share a request; reference audio requires wan3.0 or wan3.0-prime. Submit frame inputs separately. Omit for text-to-video and wan2.6 reference-to-video. Submit images in the order you want to reference them in the prompt.

Pattern: ^https://

reference_videos

(string<uri> | file)[]

Reference videos for wan2.6, wan2.7, wan3.0, and wan3.0-prime. Repeat reference_videos with publicly accessible HTTPS URLs or uploaded MP4 files with H.264 video. Use one input type within this field: all URLs or all files. For wan2.6, send 1 to 3 clips, each 2 to 30 seconds and up to 30 MB, with at least 16 fps; omit image, audio, and frame fields. For wan2.7, combine video references with image references. For wan3.0 and wan3.0-prime, send at most 5 clips totaling at most 15 seconds, with at least 16 fps. Video uploads are limited to 5 files, each at most 50 MiB (50 x 1024 x 1024 bytes); the lower wan2.6 limits still apply. Submit frame inputs separately.

Minimum array length: 1

Pattern: ^https://

reference_audios

(string<uri> | file)[]

Reference audio for wan3.0 and wan3.0-prime, combined with image or video references. Repeat reference_audios with publicly accessible HTTPS MP3 URLs or uploaded MP3 files. Use one input type within this field: all URLs or all files. Send at most 5 audio references totaling at most 15 seconds, counted separately from reference video duration. Audio uploads are limited to 5 files, each at most 15 MiB (15 x 1024 x 1024 bytes). Submit frame inputs separately. Omit for wan2.6 and wan2.7.

Minimum array length: 1

Pattern: ^https://

first_frame

Opening frame for wan2.7, wan3.0, and wan3.0-prime, as one publicly accessible HTTPS PNG or JPEG URL or one uploaded PNG or JPEG file. Use first_frame alone or add last_frame to guide the ending. Submit frame inputs in a separate request from input_reference, reference_videos, and reference_audios. Omit for wan2.6. For Wan 2.7 frame inputs, the opening image determines the aspect ratio; size selects the output resolution tier.

Pattern: ^https://

last_frame

Ending frame for wan2.7, wan3.0, and wan3.0-prime, as one publicly accessible HTTPS PNG or JPEG URL or one uploaded PNG or JPEG file. Use with first_frame to guide the ending. Submit frame inputs in a separate request from input_reference, reference_videos, and reference_audios. Omit for wan2.6. For Wan 2.7 frame inputs, the opening image determines the aspect ratio; size selects the output resolution tier.

Pattern: ^https://

响应

Task created. Store the returned id and poll GET /v1/videos/{task_id}.

string

必填

Task ID. Use this value with retrieve and content endpoints.

示例:

"task_example"

object

string

必填

Object type. Video tasks return video.

示例:

"video"

model

string

必填

Model ID used for the task.

示例:

"wan3.0"

status

enum<string>

必填

Task lifecycle status. Poll until the value is completed, failed, or error.

可用选项:

queued,

in_progress,

completed,

failed,

error

示例:

"queued"

progress

integer

必填

Task progress as a coarse percentage.

必填范围: 0 <= x <= 100

示例:

0

created_at

integer

必填

Task creation time as a Unix timestamp in seconds.

示例:

1779938152

task_id

string

Compatibility alias for id when present.

示例:

"task_example"

completed_at

integer

Task completion time as a Unix timestamp in seconds. This field appears on completed tasks.

示例:

1779938219

video_url

string

Temporary video delivery URL. This field appears on completed tasks.

示例:

"<temporary-video-url>"

error

object

Failure details. This field appears when the task fails.

最后修改于 2026年9月9日

curl https://api.cometapi.com/v1/videos \
  -H "Authorization: Bearer $COMETAPI_KEY" \
  -F model=wan3.0 \
  -F 'prompt=A small robot walks through a bright greenhouse with smooth camera movement.' \
  -F seconds=5 \
  -F size=1920x1080

{
  "id": "task_example",
  "task_id": "task_example",
  "object": "video",
  "model": "wan3.0",
  "status": "queued",
  "progress": 0,
  "created_at": 1779938152
}

​选择输入模式

​使用参考图像

​使用参考视频和音频

​Wan 2.6 视频参考

​Wan 2.7 图片和视频参考

​Wan 3.0 图片、视频和音频参考

​使用首帧和尾帧

​设置时长和尺寸

​Wan 3.0 尺寸值

​Wan 2.7 尺寸值

​Wan 2.6 尺寸值

​任务流程

授权

请求体

响应

选择输入模式

使用参考图像

使用参考视频和音频

Wan 2.6 视频参考

Wan 2.7 图片和视频参考

Wan 3.0 图片、视频和音频参考

使用首帧和尾帧

设置时长和尺寸

Wan 3.0 尺寸值

Wan 2.7 尺寸值

Wan 2.6 尺寸值

任务流程