According to AlphaMoat, as of 2026-09, SeedanceVideo has 1,168 downloads and 1 stars — ranked #4,951 of 5,078 AI skills overall, and #497 of 512 in Design & Media.
| Platform | Installs | Downloads | Stars | Comments |
|---|---|---|---|---|
| ClawHub ↗ | 26 | 1.0K | 1 | - |
| SkillHub ↗ | 161 | 1.2K | 1 | — |
Platforms count differently and update at different times.
The workflow submits text or image prompts, polls task status and retrieves an MP4 URL; the link is cleared after about 24 hours. Prompts, image URLs and task data are sent to Ark, so protect the API key. Requires curl or an equivalent reviewed HTTP tool and an Ark endpoint already deployed for video generation. Set model to that endpoint ID. Image-to-video uses one publicly reachable HTTPS image URL; the guide does not upload an arbitrary local file. Parameters depend on the chosen model. Poll for success or failure, and download using the full signed result URL including its query string before expiry.
Profile published 2026-10-27
ARK_API_KEY is configured securely.Generated from the skill’s documentation as examples; not results of actual runs.
The counting method changed in 2026-09: skills with the same name are now counted separately by author, and the full catalog is covered. Dashed lines show months before the change; they are not compared with later months, and no growth is computed across the change.
| Month | Downloads | MoM | Stars |
|---|---|---|---|
| 2026-09 | 1.2K | — | 1 |
| 2026-08 | 992 | +11.5% | 1 |
| 2026-07 | 890 | +31.5% | 1 |
| 2026-06 | 677 | +16.3% | 1 |
| 2026-05 | 582 | — | 1 |
| 2026-03 | 318 | — | 1 |
The author’s original file, shown as plain text without edits or review by AlphaMoat. Links are not clickable; check commands before running them. Fetched 2026-10-23.
---
name: Seedance Video
description: Generate AI videos with ByteDance Seedance (豆包/火山方舟) via Ark API. Supports text-to-video and image-to-video using model endpoint doubao-seedance-1-5-pro-251215. Use when the user wants to create videos from text or from an image, or mentions 火山、豆包、Seedance or AI video generation.
read_when:
- Generating video from text prompts
- Creating video from an image (image-to-video, 图生视频)
- 火山模型、豆包、Ark、Seedance video generation
metadata: {"clawdbot":{"emoji":"🎬","requires":{"env":["ARK_API_KEY"]}}}
allowed-tools: Bash(curl:*)
---
# 视频生成:火山方舟 Seedance(豆包 Ark API)
## 概述
通过**火山引擎方舟(Ark)**的 REST API 调用豆包 Seedance 视频生成模型,支持:
- **文生视频**:仅传入文本描述生成视频。
- **图生视频**:传入一张图片 + 动作/场景描述,生成视频。
- **模型**:使用推理接入点(Endpoint)ID,例如 `doubao-seedance-1-5-pro-251215`。提示词中可用参数控制时长、比例、镜头、水印等。
Base URL:`https://ark.cn-beijing.volces.com`。所有请求需在 Header 中携带:`Authorization: Bearer $ARK_API_KEY`。
## 准备
1. 在[火山方舟](https://console.volcengine.com/ark)创建 API Key,并创建/选用已部署 **Seedance 视频生成** 的推理接入点。
2. 设置环境变量(以下示例均依赖此变量):
```bash
export ARK_API_KEY="your_ark_api_key_here"
```
## 核心流程
1. **创建任务**:`POST /api/v3/contents/generations/tasks`,Body 中传 `model`(接入点 ID)和 `content`(见下)。响应返回任务 `id`。
2. **轮询状态**:`GET /api/v3/contents/generations/tasks/{id}`,直到 `status` 为 `succeeded` 或 `failed`。建议每 8–15 秒轮询一次。
3. **获取视频**:当 `status` 为 `succeeded` 时,从响应中的 `content` 里取 `video_url`(MP4,约 24 小时后清理,请及时下载)。
## 创建任务:请求体说明
- **model**(必填):推理接入点 ID,需已部署视频生成模型。例如图生视频:`doubao-seedance-1-5-pro-251215`。
- **content**(必填):数组,按顺序组合「文本」与「图片」:
- **文生视频**:一个元素即可,`type: "text"`,`text` 为视频描述,可在同一字符串内带参数,例如:`"描述内容 --duration 5 --ratio 16:9 --camerafixed false --watermark true"`。
- **图生视频**:先一个 `type: "text"`(描述动作/场景,可带 `--duration` 等参数),再一个 `type: "image_url"`,`image_url.url` 为图片公网 URL。
文本中的常用参数(以空格分隔,写在描述后面):
- `--duration 5`:时长(秒)。
- `--ratio 1:1` 或 `16:9`、`9:16`:画面比例。
- `--camerafixed true/false`:是否固定镜头(减少运动模糊)。
- `--watermark true/false`:是否加水印。
- `--fps 24`:帧率(视模型支持而定)。
## 快速开始:图生视频
```bash
# 1. 创建图生视频任务
resp=$(curl -s -X POST 'https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks' \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $ARK_API_KEY" \
-d '{
"model": "doubao-seedance-1-5-pro-251215",
"content": [
{
"type": "text",
"text": "无人机以极快速度穿越复杂障碍或自然奇观,带来沉浸式飞行体验 --duration 5 --camerafixed false --watermark true"
},
{
"type": "image_url",
"image_url": {
"url": "https://your-accessible-image-url.jpg"
}
}
]
}')
task_id=$(echo "$resp" | grep -o '"id":"[^"]*"' | cut -d'"' -f4)
echo "Task ID: $task_id"
# 2. 轮询任务状态
curl -s "https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks/$task_id" \
-H "Authorization: Bearer $ARK_API_KEY"
# 3. 当 status 为 succeeded 时,从响应的 content 中取 video_url
```
## 快速开始:文生视频
```bash
curl -s -X POST 'https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks' \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $ARK_API_KEY" \
-d '{
"model": "doubao-seedance-1-5-pro-251215",
"content": [
{
"type": "text",
"text": "戴着帽子的老爷爷面带微笑往前走 --ratio 1:1 --fps 24 --duration 5"
}
]
}'
```
## API 参考
### POST /api/v3/contents/generations/tasks(创建任务)
**Headers**
- `Content-Type: application/json`
- `Authorization: Bearer <ARK_API_KEY>`
**Body**
| 字段 | 类型 | 必填 | 说明 |
|------------|--------|------|------|
| `model` | string | 是 | 推理接入点 ID(如 doubao-seedance-1-5-pro-251215) |
| `content` | array | 是 | 内容列表,见下 |
`content` 元素:
- `type: "text"`,`text`: 视频描述 + 可选参数(如 `--duration 5 --ratio 16:9 --camerafixed false --watermark true`)。
- `type: "image_url"`,`image_url`: `{ "url": "https://..." }`,图生视频时使用,图片需公网可访问。
**响应(200)**
```json
{
"id": "cgt-2024****-**"
}
```
使用返回的 `id` 作为任务 ID 查询状态。
### GET /api/v3/contents/generations/tasks/{id}(查询任务)
**Headers**
- `Authorization: Bearer <ARK_API_KEY>`
**响应中的 status**
- `queued`:排队中
- `running`:生成中
- `succeeded`:成功,此时 `content` 中包含 `video_url`(MP4)
- `failed`:失败,可查看错误信息
成功时,视频地址在响应的 `content.video_url` 中(MP4)。视频链接约 24 小时后失效,请尽快下载。
## 示例
### 图生视频(你提供的写法)
```bash
curl -X POST https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $ARK_API_KEY" \
-d '{
"model": "doubao-seedance-1-5-pro-251215",
"content": [
{
"type": "text",
"text": "无人机以极快速度穿越复杂障碍或自然奇观,带来沉浸式飞行体验 --duration 5 --camerafixed false --watermark true"
},
{
"type": "image_url",
"image_url": {
"url": "https://ark-project.tos-cn-beijing.volces.com/doc_image/seepro_i2v.png"
}
}
]
}'
```
### 轮询并提取视频 URL
```bash
TASK_ID="cgt-2024****-**" # 替换为创建任务返回的 id
while true; do
result=$(curl -s "https://ark.cn-beijing.volces.com/api/v3/contents/generations/tasks/$TASK_ID" \
-H "Authorization: Bearer $ARK_API_KEY")
status=$(echo "$result" | grep -o '"status":"[^"]*"' | head -1 | cut -d'"' -f4)
if [ "$status" = "succeeded" ]; then
# 根据实际响应结构调整提取方式,例如:
video_url=$(echo "$result" | grep -o '"video_url":"[^"]*"' | cut -d'"' -f4)
echo "Video URL: $video_url"
break
fi
if [ "$status" = "failed" ]; then
echo "Generation failed. Check response for error details."
break
fi
sleep 10
done
```
### 下载视频
拿到 `video_url` 后建议用 curl 下载到本地再播放(链接带签名参数,浏览器复制时容易漏掉 `?` 后面的部分,会报 AuthorizationQueryParametersError):
```bash
curl -o output.mp4 "这里粘贴完整 video_url,包含 ? 及后面全部参数"
```
## 故障排查
- **401**:检查 `ARK_API_KEY` 是否正确且已导出。
- **404 / 模型不可用**:确认该接入点已部署视频生成模型,且 `model` 填的是接入点 ID(如 `doubao-seedance-1-5-pro-251215`)。
- **图生视频失败**:确保 `image_url` 为公网可访问的 HTTPS 地址,避免内网或 localhost。
- **等待时间长**:视频生成通常需 1–3 分钟,建议轮询间隔 8–15 秒。
- **链接失效**:`video_url` 约 24 小时后失效,请在成功后尽快下载。
- **打开链接显示 JSON 报错**:若看到 `AuthorizationQueryParametersError` 或 “Query-string authentication version 4 requires the X-Tos-...” 说明访问时未带完整 URL(漏了 `?` 及后面的鉴权参数)。请用 `curl -o 视频.mp4 "完整URL"` 下载到本地再播放。
- API、计费、接入点:[火山方舟控制台](https://console.volcengine.com/ark)与[视频生成 API 文档](https://www.volcengine.com/docs/82379/1520757)
- 本技能文档/示例:[GitHub Issues](https://github.com/maddy-cc/seedance-video-skill/issues)
ContextWeave one-shot architecture diagram generation
A script-based skill that generates or edits images with Nano Banana Pro.
Runs the whisper command line to transcribe audio locally.
A skill that generates and edits diagrams through an MCP server.
Design intelligence and implementation guidance for polished interfaces, from layout and flows to tokens and frontend changes.
An AI video generation and orchestration skill across several model APIs.
Pull a frame out of a video with ffmpeg, or produce a thumbnail for a quick look.
Turn audio files into text with local Whisper, running fully offline once the model is downloaded.
Data month: 2026-09 · Downloads, stars and installs are aggregated monthly from public skill registries (ClawHub, SkillHub). See methodology.