Jimeng MCP

What it does

Jimeng MCP connects your AI assistant to ByteDance's Jimeng AI platform, giving it the ability to generate high-quality images and videos on demand. It reverse-engineers the official Jimeng API, so you only need your Jimeng session cookie — no paid API key required. Supports the full Seedream image model family (jimeng-5.0 down to 1.4) and Seedance video models.

Tools

• generateImage — Generates one or more images from a text prompt. Supports model selection (jimeng-5.0 to jimeng-3.0 and legacy), custom width/height, precision (sample_strength), negative prompts, and image mixing/reference generation via a local file path or URL. Returns an array of image URLs.
• generateVideo — Generates a video clip from a text prompt. Supports first-frame and last-frame image references, resolution (720p/1080p), and all Seedance model variants. Returns a video URL.

Installation

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "jimeng": {
      "command": "node",
      "args": ["/path/to/jimeng-mcp/lib/index.js"],
      "env": {
        "JIMENG_API_TOKEN": "your_jimeng_session_id_here"
      }
    }
  }
}

Or install via Smithery for automatic setup:

npx -y @smithery/cli install @c-rick/jimeng-mcp --client claude

Getting your token: Log in at jimeng.jianying.com, open DevTools → Application → Cookies, and copy the sessionid value. Multiple tokens can be comma-separated for load balancing.

Supported hosts

Confirmed in README: Claude Desktop, Cursor.