Google Veo 3.1 MCP Server

MCP Server and CLI batch tool for Google Veo 3.1 Video Generation API.

Features

• MCP Server for integration with Claude Desktop and other MCP clients

• Batch CLI Tool for processing multiple video generation jobs

• Text-to-Video: Generate videos from text prompts

• Image-to-Video: Animate static images

• Reference Images: Use up to 3 images for character/style consistency

• Video Extension: Extend existing videos by 7 seconds

• Frame Interpolation: Generate video between two keyframes

• Cost Estimation: Calculate costs before execution

Related MCP server: veo-mcp-server

Installation

npm install
npm run build

Setup

1. Get your Google API key from Google AI Studio

2. Create a .env file:

GOOGLE_API_KEY=your_api_key_here

Or set the environment variable directly:

export GOOGLE_API_KEY=your_api_key_here

MCP Server Usage

Claude Desktop Configuration

Add to your Claude Desktop configuration (%APPDATA%\Claude\claude_desktop_config.json on Windows):

{
  "mcpServers": {
    "veo3": {
      "command": "node",
      "args": ["C:/path/to/google-veo3-1-mcp-server/dist/index.js"],
      "env": {
        "GOOGLE_API_KEY": "your_api_key_here"
      }
    }
  }
}

Or using npx after publishing:

{
  "mcpServers": {
    "veo3": {
      "command": "npx",
      "args": ["google-veo3-1-mcp-server"],
      "env": {
        "GOOGLE_API_KEY": "your_api_key_here"
      }
    }
  }
}

Available Tools

generate_video

Generate video from text prompt or image.

{
  "prompt": "A golden retriever running through autumn leaves",
  "model": "veo-3.1-generate-preview",
  "resolution": "1080p",
  "duration_seconds": 8,
  "generate_audio": true,
  "output_path": "./output/video.mp4"
}

Parameters:

• prompt: Text description of the video (required unless image provided)

• model: veo-3.1-generate-preview (default), veo-3.1-fast-generate-preview, or veo-3.1-lite-generate-preview

• aspect_ratio: 16:9 (default) or 9:16

• resolution: 720p (default), 1080p, or 4k (4k not available for lite model)

• duration_seconds: 4, 6, or 8 (default: 8; must be 8 when using 1080p/4K or reference images)

• generate_audio: Whether to generate audio (default: true)

• negative_prompt: Elements to avoid

• image: Image for Image-to-Video mode

• reference_images: Array of reference images for consistency

• sample_count: Number of videos to generate per request (default: 1)

• output_path: Path to save the video

• wait: Set to false to return the operation_name immediately instead of blocking until the video is ready (default: true). Also supported by extend_video and interpolate_frames.

extend_video

Extend an existing video by 7 seconds.

{
  "video": "./input/video.mp4",
  "prompt": "Continue with the character walking forward",
  "output_path": "./output/extended.mp4"
}

Requirements:

• Input video: 1-30 seconds, 24fps, 720p or 1080p

• Output: 7 seconds at 720p

interpolate_frames

Generate video transitioning between two keyframes.

{
  "first_frame": "./images/start.jpg",
  "last_frame": "./images/end.jpg",
  "prompt": "Smooth camera pan",
  "duration_seconds": 8,
  "generate_audio": true,
  "output_path": "./output/interpolated.mp4"
}

get_video_status

Check the status of a video generation operation. When used with wait: false generation, poll this tool until done is true, then download the result.

{
  "operation_name": "models/veo-3.1-generate-preview/operations/abc123",
  "download": true,
  "output_path": "./output/video.mp4"
}

Parameters:

• operation_name: Operation name returned by a generation tool (required)

• download: When the operation is done, download the video(s) to output_path or OUTPUT_DIR (default: false)

• output_path: Where to save the downloaded video (implies download)

Async workflow example:

1. generate_video with "wait": false → returns operation_name in seconds

2. get_video_status with the operation_name → done: false while processing

3. Once done: true, call again with "download": true (or include output_path) to save the video

Batch CLI Usage

# Estimate costs only
veo3-batch config.json --estimate-only

# Execute batch
veo3-batch config.json --output-dir ./output

# With options
veo3-batch config.json --max-concurrent 3 --no-audio --format json

CLI Options

Batch Configuration Format

{
  "jobs": [
    {
      "prompt": "A cat playing piano",
      "duration_seconds": 8,
      "resolution": "1080p",
      "generate_audio": true
    },
    {
      "type": "extend",
      "video": "./videos/source.mp4",
      "prompt": "Continue the scene"
    },
    {
      "type": "interpolate",
      "first_frame": "./images/start.jpg",
      "last_frame": "./images/end.jpg"
    }
  ],
  "output_dir": "./output",
  "max_concurrent": 2,
  "default_model": "veo-3.1-generate-preview"
}

Pricing

Prices are per second of generated video. Audio is always included.

Note: Gemini API preview models always generate audio. The generate_audio parameter is not currently supported. Use Vertex AI for audio control.

Cost Examples

Models

Reference Images

Use reference images to maintain consistency:

{
  "prompt": "The character walks through a forest",
  "reference_images": [
    {
      "image": "./character.jpg",
      "reference_type": "asset"
    }
  ]
}

• asset: For characters/objects (max 3 images)

• style: For visual style (max 1 image)

Environment Variables

API Documentation

For detailed API specifications, see:

• Google Veo 3.1 API Specification

• Gemini API Video Generation

• Vertex AI Veo Documentation

License

MIT