Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@Seedream/Seedance MCP Server Generate a 5-second video of a cat wearing a space suit on Mars"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
一个基于火山方舟（Ark）API 的 MCP 服务，提供：
Seedream 图片生成/编辑
Seedance 视频生成（异步任务）
本地 stdio 与远程 http 两种接入模式
能力 | 状态 | 说明 |
图片生成/编辑 | 可用 | 基于 |
视频生成 | 可用 | 基于 |
本地接入 | 可用 |
|
远程接入 | 可用 |
|
Related MCP server: Seedream 4.0 MCP Server
Node.js 18+
pnpm
pnpm installpnpm build变量名 | 必填 | 默认值 | 说明 |
| 是 | - | 火山方舟 API Key |
| 否 |
| 传输模式： |
| 否 |
|
|
| 否 |
|
|
| 否 |
|
|
可在项目根目录创建 .env：
ARK_API_KEY=your_ark_api_key
MCP_TRANSPORT=stdiopnpm build
node build/index.jsPowerShell 示例：
$env:ARK_API_KEY="your_ark_api_key"
$env:MCP_TRANSPORT="http"
$env:MCP_HTTP_HOST="0.0.0.0"
$env:MCP_HTTP_PORT="3000"
$env:MCP_HTTP_PATH="/mcp"
node build/index.js服务启动后地址示例：
http://127.0.0.1:3000/mcp{
"mcpServers": {
"jimenggen": {
"command": "node",
"args": [
"/path/to/jimenggen-mcp/build/index.js"
],
"env": {
"ARK_API_KEY": "your_ark_api_key",
"MCP_TRANSPORT": "stdio"
}
}
}
}不同客户端配置字段可能不同，核心是把 MCP 服务地址指向：
http://your-host:3000/mcp工具名 | 用途 |
| 文生图 |
| 图生图 |
| 通用图片生成/编辑（支持 URL 与上传文件） |
| 直接调用 Seedream API 生成图片 |
| 视频生成（异步任务，内置轮询） |
generate-digital-human
action-imitation
image-dressing
doubao-seedream-5-0-lite-260128（默认）
doubao-seedream-5-0-260128
doubao-seedream-4-5-251128
doubao-seedream-4-0-250828
doubao-seedance-1-5-pro-251215（默认）
doubao-seedance-1-0-pro-250528
doubao-seedance-1-0-pro-fast-251015
doubao-seedance-1-0-lite-t2v-250428
doubao-seedance-1-0-lite-i2v-250428
generate-video 使用异步任务流程：
创建任务：POST /contents/generations/tasks
查询任务：GET /contents/generations/tasks/{id}
成功后读取：content.video_url
任务状态：
queued
running
succeeded
failed
expired
主要参数：
参数 | 类型 | 说明 |
| string | 必填，视频提示词 |
| string | 可选，默认 |
|
| 可选，默认 |
| enum | 可选： |
| string | 可选，首帧图 URL（图生视频） |
| boolean | 可选，默认 |
| boolean | 可选，默认 |
| string | 可选，任务状态回调地址 |
| number | 可选，默认 |
| number | 可选，默认 |
兼容参数映射：
frames=121 映射为 duration=5
frames=241 映射为 duration=10
aspect_ratio 在值合法时映射到 ratio
说明：根据官方文档，Seedance 2.0 当前仅支持控制台体验，暂不支持 API 调用。
ARK_API_KEY请确认环境变量已设置，并且启动进程能读取该变量。
http 模式返回 404请检查客户端访问路径是否与 MCP_HTTP_PATH 一致，默认是 /mcp。
可适当调大 max_poll_attempts，或增大 poll_interval_ms 后重试。
Ark-only 模式下不支持该工具
Inputs: referenceAction (required), targetCharacter (required), style (required).
Ark-only 模式下不支持该工具
Inputs: prompt (required), avatarStyle (required), emotion (required), action (required).
使用火山方舟 Seedream 模型生成/编辑图片（Ark-only）
Inputs: prompt (required), ratio (required), imgUrls, uploadFiles, +3 more.
直接调用火山方舟 Seedream API 生成图片（Ark-only）
Inputs: prompt (required), model, imgUrls, size.
使用火山方舟 Seedance 模型生成视频（异步任务）
Inputs: prompt (required), model, duration, ratio, +8 more.
Ark-only 模式下不支持该工具
Inputs: modelImageUrl, garmentImageUrl.
使用火山方舟 Seedream 模型进行图生图（Ark-only）
Inputs: prompt (required), ratio (required), imageUrl, localPath, +1 more.
使用火山方舟 Seedream 模型生成图片（Ark-only）
Inputs: prompt (required), ratio (required), style, model, +3 more.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 8 tools
Multiple tools have unclear boundaries and overlapping purposes. For example, 'generate-image', 'generate-img-seedream4', and 'text-to-image' all appear to generate images, with only minor distinctions in implementation or input method that are not clearly differentiated in the descriptions. This ambiguity will likely cause misselection by agents.
The naming conventions are mixed but still readable. Most tools use a verb_noun pattern (e.g., 'generate-image', 'generate-video'), but there are deviations like 'image-dressing' and 'action-imitation' that use noun_verb or other structures. While not chaotic, the inconsistency reduces predictability.
With 8 tools, the count is reasonable and well-scoped for a media generation server focused on images and videos. It covers key operations without being overly heavy, though some tools might be redundant, which is addressed in other dimensions.
There are significant gaps in the tool surface for the domain of digital media generation. The server lacks essential operations such as editing or deleting generated content, managing tasks (e.g., checking video generation status), or handling errors. Additionally, three tools are marked as unsupported in the current mode, further limiting functionality.
LLM chat, text tools, image generation, editing, batch image jobs, and asynchronous video generation
Generate video, images, speech and music with 120+ AI models and render TSX videos to MP4
Create images and videos from prompts, with options for image mixing, reference images, and start/…
AI video generation MCP server for Kling, MiniMax, Luma & Hunyuan with prompt blueprints.