Veo 3

Generates videos from text or images using Veo 3 API with async processing and file management.

python

1

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "dayongd1-mcp-veo3": {
      "command": "uvx",
      "args": [
        "mcp-veo3",
        "--output-dir",
        "~/Videos/Generated"
      ],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here",
        "DEFAULT_OUTPUT_DIR": "generated_videos"
      }
    }
  }
}

You run an MCP server that connects to Google's Veo 3 API through Gemini to generate videos from text prompts or images. This server handles async processing, model selection, and file management so you can produce high-quality videos efficiently while controlling output settings like aspect ratio and audio.

How to use

To use your MCP Veo 3 server, start the server instance with your preferred method, then interact with it through your MCP client or orchestration layer. You can generate videos from text prompts, animate starting images, and manage the resulting files. The server exposes tools for text-to-video, image-to-video, and video information retrieval, with options to select models, set aspect ratios, and enable or disable audio depending on the model.

How to install

Prerequisites: have a Python environment and the MCP runtime tooling installed. You will also need access to Gemini API credentials for Veo 3.

Run the MCP Veo 3 server directly (no installation required) using the MCP runtime tool.
Set up a development environment if you prefer a source installation and local development workflow.
Prepare your Gemini API key for API access.

Environment and configuration

Create a configuration file to supply the API key and default settings for your environment. The following variables are used by the server.

GEMINI_API_KEY=your_gemini_api_key_here

# Optional defaults
DEFAULT_OUTPUT_DIR=generated_videos
DEFAULT_MODEL=veo-3.0-generate-preview
DEFAULT_ASPECT_RATIO=16:9
PERSON_GENERATION=dont_allow
POLL_INTERVAL=10
MAX_POLL_TIME=600

MCP client configuration options

Choose how you start and connect to the MCP Veo 3 server from your client. The following examples show explicit MCP configurations that you can paste into your client setup.

{
  "mcpServers": {
    "veo3": {
      "command": "uvx",
      "args": ["mcp-veo3", "--output-dir", "~/Videos/Generated"],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here"
      }
    }
  }
}

{
  "mcpServers": {
    "veo3": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/mcp-veo3", "mcp-veo3", "--output-dir", "~/Videos/Generated"],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here"
      }
    }
  }
}

{
  "mcpServers": {
    "veo3": {
      "command": "python",
      "args": ["/path/to/mcp-veo3/mcp_veo3.py", "--output-dir", "~/Videos/Generated"],
      "env": {
        "GEMINI_API_KEY": "your_api_key_here"
      }
    }
  }
}

Available tools

generate_video

Create a video from a text description using a specified model and optional negative prompts and aspect ratio. Returns a new video file in the output directory.

generate_video_from_image

Create a video by animating a starting image guided by a motion prompt. Supports model selection and aspect ratio customization.

list_generated_videos

List all videos generated and stored in the output directory.

get_video_info

Retrieve detailed information about a specific video file, such as resolution, duration, and file size.