マーケットサポート← アプリに戻る

Seedance 2.5

Seedance 2.5 モデルは統一された非同期ジョブ API で呼び出します。POST /api/generate でタスクを開始し、完了するまで GET /api/generate/{taskId} をポーリングします。このファミリーは 10 個のエンドポイントを提供します。

概要

すべてのエンドポイントは同じ外枠を共有し、変わるのは model と input フィールドだけです。モデルマーケットの playground から任意に実行できます。

Seedance 2.5 image-to-video modes

Three mutually exclusive modes — match fields to the model slug:

  • image-to-video: first_frame_url only; aspect_ratio must be adaptive; set first_frame_portrait_kind: skip for scenes.
  • first-last-frame: first_frame_url + last_frame_url; also requires adaptive.
  • omni-reference: reference_image_urls plus optional video/audio; mark portraits virtual/real, backgrounds skip; fixed ratios like 16:9 work.

On direct routes, only portraits are ingested to the portrait library; scenes, video, and audio stay as direct URLs. Copy the full API doc from the model API tab.

Portrait reference details
On direct routes, only portrait images are uploaded to the portrait library; scenes, videos, and audio stay as direct public URLs. Pair reference_image_urls with reference_portrait_kinds per slot:virtual (portrait library — real or AI; real is equivalent), skip (background/scene — default when omitted). Single-frame and first/last-frame models use first_frame_url / last_frame_url (ingested by default); set first_frame_portrait_kind to skip to keep a direct URL.

Seedance 2 参照メディア制限

以下は、97AI.PRO が現在接続している上流インターフェースの公開検証ルールです。公式モデルの恒久的な上限ではありません。ルートによって範囲が異なる場合があるため、API レスポンスを最終基準としてください。

参照画像

形式
JPEG, PNG, WEBP, BMP, TIFF, GIF
幅 / 高さ
300-6000 px
アスペクト比
0.4-2.5
ファイルサイズ
< 30 MB

参照動画

形式
MP4, MOV
解像度
480p, 720p
ファイルごとの長さ
2-15 s
数 / 合計時間
最大 3 個 / 合計 15 秒以内
幅 / 高さ
300-6000 px
総ピクセル数
409600-927408
アスペクト比
0.4-2.5
ファイルサイズ
<= 50 MB
フレームレート
24-60 FPS

参照音声

形式
WAV, MP3
ファイルごとの長さ
2-15 s
数 / 合計時間
最大 3 個 / 合計 15 秒以内
ファイルサイズ
<= 15 MB
サンプルレート / チャンネル
未公開
音声コーデック
未公開

出力動画

resolution
480p, 720p, 1080p, 4k
aspect_ratio
1:1, 4:3, 3:4, 16:9, 9:16, 21:9, adaptive
正確なピクセル寸法
未公開。固定の幅と高さは保証されません

認証

/api-keys で API キーを作成し、Bearer トークンとして送信します。ログイン中のブラウザからのリクエストはセッション cookie 経由でも動作します。

Authorization: Bearer <YOUR_API_KEY>

タスクの作成

POST
POST https://97ai.97claude.com/api/generate

ボディ：{ "model": "<slug>", "input": { … } }

レスポンス：{ "taskId": "<id>", "state": "waiting" }

結果の照会

GET
GET https://97ai.97claude.com/api/generate/<taskId>

レスポンス：{ "taskId", "state": "waiting" | "success" | "fail", "resultUrls": [ … ], "creditsConsumed", "failMsg", "balance" }

非同期
タスク作成の 200 は、タスクが受理されたことを意味するだけで、完了ではありません。state が success または fail になるまでこのエンドポイントをポーリングしてください。失敗したタスクは課金されません。

エンドポイント

モデルタスクmodel slug
Seedance 2.5Text to Videoseedance-2.5-text-to-video
Seedance 2.5 Text to VideoText to Videoseedance-2.5-text-to-video-480p
Seedance 2.5 Image to VideoImage to Videoseedance-2.5-image-to-video
Seedance 2.5 Image to VideoImage to Videoseedance-2.5-image-to-video-480p
Seedance 2.5 First & Last FrameImage to Videoseedance-2.5-first-last-frame
Seedance 2.5 First & Last FrameImage to Videoseedance-2.5-first-last-frame-480p
Seedance 2.5 Omni ReferenceImage to Videoseedance-2.5-omni-reference
Seedance 2.5 Omni ReferenceImage to Videoseedance-2.5-omni-reference-480p
Seedance 2.5 Spicy Text to VideoText to Videoseedance-2.5-spicy-text-to-video
Seedance 2.5 Spicy Image to VideoImage to Videoseedance-2.5-spicy-image-to-video

Seedance 2.5 · Text to Video

POSTText to Videoseedance-2.5-text-to-video
input フィールド
first_frame_urlstring任意

First-frame image URL or asset://{assetId}. Mutually exclusive with reference_* fields.

last_frame_urlstring任意

Last-frame image URL or asset://{assetId}. Requires first_frame_url.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

last_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for last_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the video to generate (max 30000 characters).

reference_image_urlsarray<string>任意

Reference image URLs (max 30). Mutually exclusive with first/last frame fields.

reference_video_urlsarray<string>任意

Reference video URLs (max 10; total duration ≤30s).

reference_audio_urlsarray<string>任意

Reference audio URLs (max 10; total duration ≤30s).

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Output aspect ratio.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-text-to-video",
  "input": {
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞"
  }
}'

Seedance 2.5 Text to Video · Text to Video

POSTText to Videoseedance-2.5-text-to-video
input フィールド
first_frame_urlstring任意

First-frame image URL or asset://{assetId}. Mutually exclusive with reference_* fields.

last_frame_urlstring任意

Last-frame image URL or asset://{assetId}. Requires first_frame_url.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

last_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for last_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the video to generate (max 30000 characters).

reference_image_urlsarray<string>任意

Reference image URLs (max 30). Mutually exclusive with first/last frame fields.

reference_video_urlsarray<string>任意

Reference video URLs (max 10; total duration ≤30s).

reference_audio_urlsarray<string>任意

Reference audio URLs (max 10; total duration ≤30s).

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Output aspect ratio.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-text-to-video",
  "input": {
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "480p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞"
  }
}'

Seedance 2.5 Image to Video · Image to Video

POSTImage to Videoseedance-2.5-image-to-video
input フィールド
first_frame_urlstring必須

URL of the input image to animate, or asset://{assetId} after portrait ingest.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the video motion and style.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Must be adaptive — the direct route follows the input image aspect ratio for keyframe-driven tasks.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-image-to-video",
  "input": {
    "aspect_ratio": "adaptive",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞",
    "first_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
    "first_frame_portrait_kind": "virtual"
  }
}'

Seedance 2.5 Image to Video · Image to Video

POSTImage to Videoseedance-2.5-image-to-video-480p
input フィールド
first_frame_urlstring必須

URL of the input image to animate, or asset://{assetId} after portrait ingest.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the video motion and style.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Must be adaptive — the direct route follows the input image aspect ratio for keyframe-driven tasks.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-image-to-video-480p",
  "input": {
    "aspect_ratio": "adaptive",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "480p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞",
    "first_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
    "first_frame_portrait_kind": "virtual"
  }
}'

Seedance 2.5 First & Last Frame · Image to Video

POSTImage to Videoseedance-2.5-first-last-frame
input フィールド
first_frame_urlstring必須

First-frame image URL or asset://{assetId}. Mutually exclusive with reference_* fields.

last_frame_urlstring必須

Last-frame image URL or asset://{assetId}. Requires first_frame_url.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

last_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for last_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the transition and motion.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Must be adaptive — the direct route follows the input image aspect ratio for keyframe-driven tasks.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-first-last-frame",
  "input": {
    "aspect_ratio": "adaptive",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞",
    "first_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
    "first_frame_portrait_kind": "virtual",
    "last_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/image/20260819/fb280368-8f54-41eb-aacf-cee1c189b29b.png",
    "last_frame_portrait_kind": "skip"
  }
}'

Seedance 2.5 First & Last Frame · Image to Video

POSTImage to Videoseedance-2.5-first-last-frame-480p
input フィールド
first_frame_urlstring必須

First-frame image URL or asset://{assetId}. Mutually exclusive with reference_* fields.

last_frame_urlstring必須

Last-frame image URL or asset://{assetId}. Requires first_frame_url.

first_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for first_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

last_frame_portrait_kindenum<string>任意

Portrait-library route: ingest hint for last_frame_url. Default skip (direct URL). Use virtual/real to upload portraits to the asset library.

promptstring必須

Text prompt describing the transition and motion.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Must be adaptive — the direct route follows the input image aspect ratio for keyframe-driven tasks.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-first-last-frame-480p",
  "input": {
    "aspect_ratio": "adaptive",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "480p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞",
    "first_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
    "first_frame_portrait_kind": "virtual",
    "last_frame_url": "https://frw-images.aiaipic.win/users/3514640099037941760/image/20260819/fb280368-8f54-41eb-aacf-cee1c189b29b.png",
    "last_frame_portrait_kind": "skip"
  }
}'

Seedance 2.5 Omni Reference · Image to Video

POSTImage to Videoseedance-2.5-omni-reference
input フィールド
promptstring必須

Prompt describing the target video, style, dynamics, and atmosphere.

reference_image_urlsarray<string>任意

Reference image URLs (max 30). Scene/background images stay as URLs unless marked for portrait ingest.

reference_portrait_kindsarray<string>任意

Per-image ingest hint aligned with reference_image_urls: virtual (portrait library — AI or real person), real (optional alias, same as virtual), skip (direct URL; default when omitted).

reference_imagesarray<object>任意

Structured reference images: [{ url, portrait: "virtual"|"real"|"skip" }]. Alternative to reference_image_urls + reference_portrait_kinds.

virtual_portraitsarray<string>任意

Portrait URLs always uploaded to the virtual portrait library before generation.

real_portraitsarray<string>任意

Portrait URLs (including real-person photos) uploaded to the virtual portrait library before generation.

reference_video_urlsarray<string>任意

Reference video URLs (max 10; total duration ≤30s). Never uploaded to the asset library.

reference_audio_urlsarray<string>任意

Reference audio URLs (max 10; total duration ≤30s). Never uploaded to the asset library.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Output aspect ratio.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-omni-reference",
  "input": {
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": " @image1与 @image2在 @image3的场景中，参考 @video1的运镜，背景音乐参考 @audio1",
    "reference_image_urls": [
      "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
      "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
      "https://frw-images.aiaipic.win/users/3514640099037941760/image/20260819/fb280368-8f54-41eb-aacf-cee1c189b29b.png"
    ],
    "reference_portrait_kinds": [
      "virtual",
      "virtual",
      "skip"
    ],
    "reference_video_urls": [
      "https://frw-images.aiaipic.win/users/3492453287637028864/video/20260818/3524765235086888960.mp4"
    ],
    "reference_audio_urls": [
      "https://frw-images.aiaipic.win/users/3492453287637028864/audio/20260818/3524725649950380032.mp3"
    ]
  }
}'

Seedance 2.5 Omni Reference · Image to Video

POSTImage to Videoseedance-2.5-omni-reference-480p
input フィールド
promptstring必須

Prompt describing the target video, style, dynamics, and atmosphere.

reference_image_urlsarray<string>任意

Reference image URLs (max 30). Scene/background images stay as URLs unless marked for portrait ingest.

reference_portrait_kindsarray<string>任意

Per-image ingest hint aligned with reference_image_urls: virtual (portrait library — AI or real person), real (optional alias, same as virtual), skip (direct URL; default when omitted).

reference_imagesarray<object>任意

Structured reference images: [{ url, portrait: "virtual"|"real"|"skip" }]. Alternative to reference_image_urls + reference_portrait_kinds.

virtual_portraitsarray<string>任意

Portrait URLs always uploaded to the virtual portrait library before generation.

real_portraitsarray<string>任意

Portrait URLs (including real-person photos) uploaded to the virtual portrait library before generation.

reference_video_urlsarray<string>任意

Reference video URLs (max 10; total duration ≤30s). Never uploaded to the asset library.

reference_audio_urlsarray<string>任意

Reference audio URLs (max 10; total duration ≤30s). Never uploaded to the asset library.

generate_audioboolean任意

Generate synchronized audio with the video.

return_last_frameboolean任意

Return the generated video's last frame. Cannot be true when draft=true.

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Output aspect ratio.

durationinteger任意

Video duration from 4 to 30 seconds, or -1 for automatic selection.

output_formatenum<string>任意

Video output format.

web_searchboolean任意

Enable online search for the generation.

nsfw_checkerboolean任意

Enable content safety checking.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-omni-reference-480p",
  "input": {
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "480p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": " @image1与 @image2在 @image3的场景中，参考 @video1的运镜，背景音乐参考 @audio1",
    "reference_image_urls": [
      "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
      "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
      "https://frw-images.aiaipic.win/users/3514640099037941760/image/20260819/fb280368-8f54-41eb-aacf-cee1c189b29b.png"
    ],
    "reference_portrait_kinds": [
      "virtual",
      "virtual",
      "skip"
    ],
    "reference_video_urls": [
      "https://frw-images.aiaipic.win/users/3492453287637028864/video/20260818/3524765235086888960.mp4"
    ],
    "reference_audio_urls": [
      "https://frw-images.aiaipic.win/users/3492453287637028864/audio/20260818/3524725649950380032.mp3"
    ]
  }
}'

Seedance 2.5 Spicy Text to Video · Text to Video

POSTText to Videoseedance-2.5-spicy-text-to-video
input フィールド
promptstring必須

Text prompt describing the video motion and style. Spicy mode favors bolder, higher-contrast results.

durationinteger任意

Video length in seconds (4-16).

resolutionenum<string>任意

Output video resolution.

aspect_ratioenum<string>任意

Output video aspect ratio.

camera_fixedboolean任意

Whether to keep the camera fixed.

generate_audioboolean任意

Whether to generate synchronized audio.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-spicy-text-to-video",
  "input": {
    "aspect_ratio": "16:9",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞"
  }
}'

Seedance 2.5 Spicy Image to Video · Image to Video

POSTImage to Videoseedance-2.5-spicy-image-to-video
input フィールド
promptstring必須

Text prompt describing the video motion and style. Spicy mode favors bolder, higher-contrast results.

image_urlstring必須

URL of the input image to animate.

last_imagestring任意

Optional closing image. Together with image_url, this selects strict first/last-frame keyframe generation.

durationinteger任意

Normal mode: 4-16 seconds. Two-image keyframe mode: 4-30 seconds.

resolutionenum<string>任意

Normal mode supports its listed tiers. Two-image keyframe mode supports 480p, 720p, or 1080p.

aspect_ratioenum<string>任意

Normal mode uses a selected ratio. Two-image keyframe mode requires adaptive and follows the opening image ratio.

camera_fixedboolean任意

Whether to keep the camera fixed.

generate_audioboolean任意

Whether to generate synchronized audio.

watermarkboolean任意

For two-image keyframe mode, add an AI-generated watermark to the output.

例
cURL
curl -X POST https://97ai.97claude.com/api/generate \
  -H "Authorization: Bearer $API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "seedance-2.5-spicy-image-to-video",
  "input": {
    "aspect_ratio": "adaptive",
    "duration": 5,
    "generate_audio": true,
    "output_format": "mp4",
    "resolution": "720p",
    "return_last_frame": false,
    "web_search": false,
    "prompt": "在房间里跳舞",
    "image_url": "https://frw-images.aiaipic.win/users/3514640099037941760/ai_generated/20260818/038f9f68-004f-4017-a09d-74b3da84e569.png",
    "first_frame_portrait_kind": "virtual"
  }
}'