Doubao Image/Video

You can run this MCP server to generate images with the Seedream model and videos with the Seedance model, plus check the status of video tasks. It centralizes image and video inference behind a single MCP client, making it easy to integrate generation capabilities into your apps.

How to use

Use an MCP client to connect to the Doubao Image/Video MCP Server. You can request single image generations or start asynchronous video generations. You can also query the status of video tasks to see when a video is ready.

How to install

Prerequisites: you need Node.js and a package manager. The recommended flow uses the MCP client via the npx command.

cd doubao-image-video-mcp
pnpm install
pnpm build

Configuration examples

{
  "mcpServers": {
    "doubao-giv": {
      "command": "npx",
      "args": ["-y", "doubao-image-video-mcp@latest"],
      "env": {
        "DOUBAO_API_KEY": "your_api_key_here",
        "DOUBAO_IMAGE_ENDPOINT_ID": "ep-20241227-xxxxxxxxxxxxx",
        "DOUBAO_VIDEO_ENDPOINT_ID": "ep-20241227-xxxxxxxxxxxxx"
      }
    }
  }
}

Environment variables

The following environment variables are used to run the MCP server. Provide your own values when configuring the client.

DOUBAO_API_KEY: Your API key for authenticating with the Doubao services
DOUBAO_IMAGE_ENDPOINT_ID: Endpoint ID for image generation access
DOUBAO_VIDEO_ENDPOINT_ID: Endpoint ID for video generation access
DOUBAO_DEFAULT_IMAGE_MODEL: Optional default image model (e.g., doubao-seedream-4-5)
DOUBAO_DEFAULT_VIDEO_MODEL: Optional default video model (e.g., doubao-seedance-1.0-lite-t2v)

HTTP and local (stdio) MCP configurations

The server supports a local, stdio-based MCP client configuration. Use the following example to start the MCP client with the required environment variables.

{
  "mcpServers": {
    "doubao-giv": {
      "command": "npx",
      "args": ["-y", "doubao-image-video-mcp@latest"],
      "env": {
        "DOUBAO_API_KEY": "your_api_key_here",
        "DOUBAO_IMAGE_ENDPOINT_ID": "ep-20241227-xxxxxxxxxxxxx",
        "DOUBAO_VIDEO_ENDPOINT_ID": "ep-20241227-xxxxxxxxxxxxx"
      }
    }
  }
}

Supported tools

This MCP server provides three main capabilities for you to use in your applications.

- generate_image: Create a single image from a textual prompt using the Seedream model. Supports optional image references and multiple image sizes.
- generate_video: Launch an asynchronous video generation task using the Seedance model. Returns a task ID for later status checks.
- query_video_task: Check the status and result of a previously started video generation task.

Available tools

generate_image

Generate a single image from a text prompt. Supports model selection, image size, reference images, and multi-image fusion.

generate_video

Start an asynchronous video generation job from a text prompt. Specify model, duration, resolution, and other options. Returns a task ID for status tracking.

query_video_task

Query the status and result of a previously started video generation task using its task ID.