在 OpenClaw 中使用 Seedance。
Seedance 2.0 可生成 4–15 秒、最高 1080p 的视频，按秒计费起价约 $0.05/秒——是 RunAPI 视频模型中每秒成本最低的。它支持多模态参考输入（图像、视频和音频 URL）、首/末帧控制以及原生音频同步。OpenClaw agent 通过与聊天相同的 RunAPI 密钥和端点调用它。
通过 RunAPI 使用 Seedance 2.0 生成产品展示视频。
要求：
- 从 RUNAPI_API_KEY 读取 API 密钥。
- 调用 POST https://runapi.ai/api/v1/seedance/text_to_video
- 将 model 设置为 "seedance-2.0"。
- 将 prompt 设置为描述视频场景。
- 将 duration_seconds 设置为所需时长（4–15）。
- 将 aspect_ratio 设置为 "16:9" 以生成横向视频。
- 任务是异步的。轮询返回的 task_id 直到 status 为 "completed"。
- 完成后，从响应 output 中读取视频 URL。
curl -X POST https://runapi.ai/api/v1/seedance/text_to_video \
-H "Authorization: Bearer $RUNAPI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "seedance-2.0",
"prompt": "A ceramic coffee mug rotates slowly on a wooden table, steam rising, warm morning light from a window, cinematic",
"duration_seconds": 8,
"aspect_ratio": "16:9",
"output_resolution": "1080p"
}'
{
"task_id": "tsk_abc123",
"status": "pending",
"model": "seedance-2.0"
}
三步在 OpenClaw 中使用 Seedance
Configure RunAPI
Set RUNAPI_API_KEY in your environment. If you already configured RunAPI for chat in OpenClaw, the same key works for Seedance video generation — no extra provider setup needed.
export RUNAPI_API_KEY=runapi_xxx
Call Seedance text_to_video
Send a POST to the text_to_video endpoint with model set to seedance-2.0. Include a prompt, duration_seconds (4–15), aspect_ratio, and output_resolution. Optionally attach first_frame_image_url, reference_image_urls, or reference_audio_urls for multimodal control.
POST /api/v1/seedance/text_to_video
Poll for the result
The endpoint returns a task_id immediately. Poll the task status endpoint until the status is completed, then read the output video URL from the response. Billing is per-second based on the generated video duration and resolution.
GET /api/v1/seedance/text_to_video/tsk_abc123
Seedance API 参数
| 参数 | 类型 | 说明 |
|---|---|---|
model |
string |
Required. seedance-2.0, seedance-2.0-fast, seedance-1.5-pro, seedance-v1-pro, seedance-v1-lite. |
prompt |
string |
Required. Text description of the video scene to generate. |
duration_seconds |
integer |
Video length in seconds. 4–15 for Seedance 2.0, 5 or 10 for v1 variants. |
aspect_ratio |
string |
Output aspect ratio. Accepted: 1:1, 4:3, 3:4, 16:9, 9:16, 21:9, auto. |
output_resolution |
string |
Video resolution. Accepted: 480p, 720p, 1080p. Higher resolution costs more per second. |
first_frame_image_url |
string |
Optional. URL of an image to use as the first frame. Enables image-to-video workflows. |
last_frame_image_url |
string |
Optional. URL of an image to use as the last frame. Constrains the ending scene. |
reference_image_urls |
array |
Optional. URLs of reference images for pose, depth, or edge guidance. |
reference_audio_urls |
array |
Optional. URLs of reference audio files for native audio sync in the generated video. |
reference_video_urls |
array |
Optional. URLs of reference videos for motion or style transfer. |
callback_url |
string |
Optional. Webhook URL that receives a POST when the task completes. |
OpenClaw 上的 Seedance 是什么？
字节跳动的 Seedance 2.0 是排行榜领先的视频模型，具备原生音频集成和精准的音视频同步能力。它生成 4 至 15 秒、最高 1080p 的视频，支持多镜头叙事、导演级场景转场控制和真正可用的口型同步功能。OpenClaw agent 通过 RunAPI 以按秒计费的方式调用它。
Seedance 使用场景
对话密集的短片
生成角色说话的短片——Seedance 能处理跨多个说话者的口型同步和对白归属，适合对话内容和说话头像场景。
TikTok 和 Reels 短视频
生成精确时长的平台适配竖版视频。按秒计费意味着 5 秒短片的成本约为 15 秒视频的三分之一。
产品演示样片
创作带音频旁白与视觉动作同步的产品演示短片。传入参考音频，将视频动作对齐旁白音轨。
Seedance + OpenClaw 常见问题
Yes. Configure RunAPI as an OpenAI-compatible provider in OpenClaw, then call the text_to_video endpoint with model set to seedance-2.0. The same API key handles both chat and video generation.
You are billed based on the actual duration of the generated video and the output resolution you select. A 480p video costs less per second than 1080p. The cost is deducted after the video completes, not when the task is created.
Yes. Seedance accepts English prompts and generates multi-language content. It supports both English and Chinese prompt input regardless of the target video content language.
Yes. Pass an image URL in the first_frame_image_url parameter along with a text prompt. Seedance uses the image as the opening frame and generates motion from there. You can also set last_frame_image_url to constrain the ending.
Seedance 2.0 can sync generated video motion to an audio track. Pass the audio file URL in reference_audio_urls and the model aligns visual motion to the audio -- useful for music videos or dialogue-driven scenes without a separate audio-to-video step.