Seedance 2.5 API
由文字或图片生成电影感视频。
- 4 到 30 秒的片段
- 先出 480p 样片确认镜头,再渲染正式片
| 分辨率 | 文本或图片 | 含视频输入 |
|---|---|---|
| 480p | 文本或图片26.4 积分/秒 ($0.132/秒)比 fal 便宜 40% fal 价格 $0.221/秒 | 含视频输入15.8 积分/秒 ($0.079/秒) |
| 720p | 文本或图片56.6 积分/秒 ($0.283/秒)比 fal 便宜 40% fal 价格 $0.473/秒 | 含视频输入34 积分/秒 ($0.17/秒) |
| 1080p | 文本或图片136.4 积分/秒 ($0.682/秒)比 fal 便宜 40% fal 价格 $1.137/秒 | 含视频输入81.8 积分/秒 ($0.409/秒) |
含视频输入时,输入和输出的秒数都计费。
完整价格 ↓全部 Seedance 模型价格
价格自 2026-11-09 起生效
| 档位 | 价格 |
|---|---|
480p | 26.4 积分/秒 ($0.132/秒)注 14–30 秒 |
720p | 56.6 积分/秒 ($0.283/秒)注 14–30 秒 |
1080p | 136.4 积分/秒 ($0.682/秒)注 14–30 秒 |
480p(含视频输入) 480p-video | 15.8 积分/秒 ($0.079/秒)注 2, 34–30 秒 |
720p(含视频输入) 720p-video | 34 积分/秒 ($0.17/秒)注 2, 34–30 秒 |
1080p(含视频输入) 1080p-video | 81.8 积分/秒 ($0.409/秒)注 2, 34–30 秒 |
- 最多冻结 30 秒,按实际时长扣费。
- 费率同时适用于输入和输出的秒数。
- 最多冻结 30 秒输出加 30 秒输入,按实际时长扣费。
1 积分 ≈ $0.005。按标准充值包计算;较大的充值包含赠送积分,实际单价更低。
720p:fal $0.473/秒,VizoAPI $0.283/秒,低 40%。
交付与计费
- 成品
- 成品文件列在任务结果中,每个文件有素材 ID 和类型。
- 下载
- 文件保存 7 天,通过短期签名链接下载,可随时删除。
- 计费
- 任务受理时冻结积分,交付后扣费。生成失败,或 24 小时内未交付,冻结的积分会全部释放。
- 通知
- 按任务 ID 轮询,或使用签名 Webhook,并按 event_id 去重。
如何计费?
生成失败会扣费吗?
API 参考
任务只创建一次,之后按 ID 查询。 完整脚本(含上传和下载) · API 文档
API=https://api.vizoapi.com/v1
AUTH="Authorization: Bearer $VIZO_API_KEY"
# Create the task once and keep its ID.
ID=$(curl -fsS $API/generations -H "$AUTH" -H "Content-Type: application/json" \
-d '{
"model": "seedance-2-5",
"input": {
"content": [
{
"type": "text",
"text": "一段 12 秒的电影感太空歌剧,共四个镜头,史诗而安静,胶片颗粒,自然光。\n镜头 1:远古草原的黎明,广角低机位。几个原始人类警惕地蹲伏在参考图中那块高大的黑色石碑周围。光滑的碑面上刻着一个开口圆环,缺口在右上方。\n镜头 2:镜头缓慢推向石碑。琥珀色的太阳从石碑背后升起,正好嵌进圆环的缺口,补全了这个符号。一个原始人站起来,举起一个带着长长弯角、被晒白的动物头骨,长啸一声,把它奋力抛向高空。\n镜头 3:慢动作,低机位仰拍,带角的头骨在黎明的蓝天里高高翻滚,两只弯角连成一道开口的圆弧。匹配剪辑:翻滚的头骨变成参考图中那座深灰色的环形空间站,在漆黑的太空中以同样的方式翻滚。\n镜头 4:太空中的固定机位。空间站慢慢停下,正对镜头:一个开口圆环,缺口在右上方,环上有一排细小的灯光。一颗琥珀色的行星从它后方升起,正好填满缺口。画面定格。\n声音:风声,远处的鸟叫,低沉的吼声,渐强的低频嗡鸣,一声长啸,头骨划破空气的呼啸,然后是太空的寂静和一段悠长低沉的合成器长音。没有配乐,没有人声,没有文字。"
}
],
"resolution": "720p",
"ratio": "16:9",
"duration": 5
},
"metadata": {
"client_reference": "my-unique-reference-1"
}
}' | jq -r .id)
# Poll until the task closes, then read its status and result.
until curl -fsS $API/generations/$ID -H "$AUTH" | jq -e '.closed_at != null' > /dev/null; do sleep 5; done
curl -fsS $API/generations/$ID -H "$AUTH" | jq '{status, result}'全部字段
在 input 中发送这些字段。表中没有的字段会被拒绝,返回 HTTP 422 invalid_input。
- Apart from a
draft_taskrequest,contentneeds at least one text, image or video item. - Every image, video and audio item in
contentnames itsrole. - Input limits: up to 30 images, 10 videos and 10 audio items in
content, with no limit on the total. Each video and audio item runs 2 to 30 seconds, and aneditsource video 4 to 30 seconds. Videos add up to at most 30 seconds, and audio to at most 30 seconds. Lengths are checked before the task is created when the file can be measured (422input_duration_out_of_range). - Image pixel size and aspect ratio, and video resolution and frame rate, are not checked when the request is made: the model checks them when the task runs, and a file it refuses usually fails the task with
input_media_invalid(otherwisegeneration_failed); the task is not charged. first_frameandlast_frameimages (at most one each) cannot be combined with reference images, videos or audio, and alast_frameneeds afirst_frame.ratiomust beadaptivefor first-frame, edit and extend input.omni_reference_task_typemay be omitted or auto with a reference video; it is not allowed without video. Automatic task selection holds the maximum output length to cover edit or extend; foredit,durationmust be-1.draftset totrueneedsresolution480pand cannot be combined withreturn_last_frame.- A
draft_taskitem must be the only item incontent, and names a succeeded draft of your project from the last 7 days. - A
draft_taskrequest reuses the draft's settings: it does not acceptduration,ratio,seed,generate_audio,omni_reference_task_type,draftorbitrate_mode, andresolutioncan only be1080p. - Prompt text must not contain inline
--flagparameters.
| 字段 | 说明 |
|---|---|
| Prompt text and media items. Each image, video and audio item names its role. 至少 1 项 |
content[]. | Item kind. 可选值 textimage_urlvideo_urlaudio_urldraft_task |
type 为 text 的项 | |
content[]. | Prompt text, up to 4000 characters per item. 最多 4000 个字符 |
type 为 image_url 的项 | |
content[]. | Image roles: first_frame, last_frame, reference_image. Video: reference_video. Audio: reference_audio. |
content[]. | |
content[]. | An https URL or an uploaded asset id (ast_...). 最多 4000 个字符 · 格式 |
type 为 video_url 的项 | |
content[]. | Image roles: first_frame, last_frame, reference_image. Video: reference_video. Audio: reference_audio. |
content[]. | |
content[]. | An https URL or an uploaded asset id (ast_...). 最多 4000 个字符 · 格式 |
type 为 audio_url 的项 | |
content[]. | Image roles: first_frame, last_frame, reference_image. Video: reference_video. Audio: reference_audio. |
content[]. | |
content[]. | An https URL or an uploaded asset id (ast_...). 最多 4000 个字符 · 格式 |
type 为 draft_task 的项 | |
content[]. | The finished draft to upgrade. Must be the only content item. |
content[]. | Generation id of a succeeded draft (gen_...). 格式 |
| Output resolution. Defaults to 720p, or 1080p when the content is a draft. 可选值 480p720p1080p默认值 |
| Output aspect ratio. Must be adaptive for first-frame, edit and extend inputs. Defaults to adaptive. 可选值 16:94:31:13:49:1621:9adaptive默认值 |
| Output length in seconds, 4 to 30, or -1 to let the model choose. Defaults to 5. Must be -1 for edit. 默认值 4 到 30,或 |
| Adds a synchronized soundtrack. Defaults to true. 默认值 |
| Random seed. -1 or omitted picks one at random. 最小 -1 · 最大 4,294,967,295 |
| Also returns the last frame of the video as an image. |
| Container format. Only mp4. 固定值 |
| Encoding quality. Only standard. 固定值 |
| Generates a quick 480p preview that a later request can upgrade to 1080p. |
| With a reference video, omit or choose auto to infer the task. Reference makes a new video, edit changes the source, and extend lengthens it. Automatic selection holds the maximum output length. 可选值 referenceeditextendauto |
| Add a visible watermark. Defaults to false; the tier price is unchanged. 默认值 |
| Execution and delivery deadline in seconds from acceptance. Defaults to 86400. Credits remain held until completion or this deadline. 默认值 最小 3600 · 最大 259,200 |
完整请求与响应示例
完整的创建请求、202 响应,以及任务成功后的状态。 此输入选择 720p 档位,每秒 56.6 积分。5 秒共冻结 283.000 积分。
请求
{
"model": "seedance-2-5",
"input": {
"content": [
{
"type": "text",
"text": "A paper boat drifting on a pond at dawn, slow dolly in"
}
],
"resolution": "720p",
"ratio": "16:9",
"duration": 5
},
"metadata": {
"client_reference": "my-unique-reference-1"
}
}202 已受理
{
"id": "gen_01JD8Z5QW3XR",
"model": "seedance-2-5",
"source": "api",
"created_at": "2026-11-08T09:00:00Z",
"customer_resolution_deadline": "2026-11-09T09:00:00Z",
"late_result_available": false,
"metadata": {
"client_reference": "my-unique-reference-1"
},
"status": "processing",
"execution_status": "processing",
"delivery_status": "pending",
"billing_status": "pending",
"version": 1,
"price": {
"unit": "credit",
"estimated": "283.000",
"reserved": "0.000",
"charged": "0.000",
"refunded": "0.000"
},
"allowed_actions": [],
"closed_at": null,
"error": null,
"result": null
}200 已成功
{
"id": "gen_01JD8Z5QW3XR",
"model": "seedance-2-5",
"source": "api",
"created_at": "2026-11-08T09:00:00Z",
"customer_resolution_deadline": "2026-11-09T09:00:00Z",
"late_result_available": false,
"metadata": {
"client_reference": "my-unique-reference-1"
},
"status": "succeeded",
"execution_status": "succeeded",
"delivery_status": "ready",
"billing_status": "settled",
"version": 4,
"price": {
"unit": "credit",
"estimated": "283.000",
"reserved": "0.000",
"charged": "283.000",
"refunded": "0.000"
},
"allowed_actions": [],
"closed_at": "2026-11-08T09:01:31Z",
"error": null,
"result": {
"asset_id": "ast_01JD8Z9M2C4V",
"outputs": [
{
"index": 0,
"asset_id": "ast_01JD8Z9M2C4V",
"content_type": "video/mp4"
}
]
}
}

