​
Features

  • Generate video from image or text prompt
  • Get generation task status
  • Automatically download the generated video
Note: Before using the service, you need to have the Pollo AI API key.

​
Installation

npm install -g pollo-mcp

​
Quick Start

Pollo MCP Server implements the Model Context Protocol (MCP) specification and can be used as a server to interact with MCP-compatible clients (such as Cursor).

​
Quickstart with Cursor

  1. Get your API key from Pollo AI API platform.
  2. Make sure that you already installed Node.js and npm.
Go to Cursor → Preferences → Cursor Settings → MCP Tools → New MCP Server, add config. 
{
  "mcpServers": {
    "pollo.ai": {
      "command": "npx",
      "args": ["-y", "mcp-server-pollo"],
      "env": {
        "POLLO_AI_BASE_URL": "https://pollo.ai",
        "POLLO_AI_API_KEY": "<your-api-key>",
        "POLLO_AI_HOME_DIR": "<path to save generated videos>",
        "POLLO_AI_VIDEO_MODEL_IMG": "<Model for generating videos from images, optional, pollo-v1-6 is used by default, multiple models are separated by commas>",
        "POLLO_AI_VIDEO_MODEL_TEXT": "<Model for generating videos from text, optional, pollo-v1-6 is used by default, multiple models are separated by commas>"
      }
    }
  }
}
That’s it. Your MCP client can now interact with Pollo.ai through these tools.

​
Tools

  • text2video
    • Generate video from text prompt.
    • Input: model(object)，supports multiple models, for a detailed list, see Pollo AI API Platform
      • modelBrand(required, string), e.g., pollo, kling-ai, runway.
      • modelAlias(required, string), e.g., pollo-v1-6, kling-v2, runway-gen-4-turbo.
      • prompt(required, string), the prompt of the generation
      • other model supported config, see Pollo AI API Platform
  • img2video
    • Generate video from image.
    • Input: model(object)，supports multiple models, for a detailed list, see Pollo AI API Platform
      • modelBrand(required, string), e.g., pollo, kling-ai, runway.
      • modelAlias(required, string), e.g., pollo-v1-6, kling-v2, runway-gen-4-turbo.
      • any of
        • imagePath(string) and imageTailPath(string)，the local image path used to generate the video.
        • image(string) and imageTail(string), the image url used to generate the video. Only image URLs are accepted; base64 is not allowed.
      • other model supported config, see Pollo AI API Platform
  • getTaskStatus
    • Check generation task status, and automatically download the completed video.
    • Inputs:
      • taskId (string): generation task id returned by text2video and img2video