Luma AI

MCP server for Luma AI Dream Machine API enabling video/image generation, keyframe tooling, upscaling, audio, and credits management.

python

5

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": {
    "bobtista-luma-ai-mcp-server": {
      "command": "uv",
      "args": [
        "run",
        "--project",
        "/path/to/your/luma-ai-mcp-server",
        "-m",
        "luma_ai_mcp_server"
      ],
      "env": {
        "LUMA_API_KEY": "YOUR_LUMA_API_KEY"
      }
    }
  }
}

You can run a dedicated MCP server for Luma AI's Dream Machine API to generate and manage AI-generated videos and images, control their lifecycle through a unified interface, and integrate those capabilities with your favorite MCP clients.

How to use

You interact with the Luma AI MCP Server by sending MCP-style tool calls to the server through a suitable MCP client. Use it to create new video generations from prompts, monitor their status, and perform enhancements such as upscaling or adding AI-generated audio. You can also generate images from prompts, retrieve your current credit balance, and access a set of camera motion options. Leverage keyframes for advanced video generation and manage generations with listing, getting status, and deletion operations.

How to install

Prerequisites: you need a working environment where you can run MCP servers and supply environment variables. You will also need an API key from Luma AI for the Dream Machine API.

Prepare the runtime for the MCP server. The server configuration shown uses a local runner that executes the MCP server project with a runtime helper.
Create the configuration file for your MCP client, including the Luma API key and the command to start the MCP server.
Start the MCP server using the runtime helper with your project path and module name as shown in the example configuration.

Additional content

Tools exposed by this MCP server include ping, create_generation, get_generation, list_generations, delete_generation, upscale_generation, add_audio, generate_image, get_credits, and get_camera_motions. Each tool has specific inputs and outputs described in the tool definitions. The server supports a range of video and image generation options, including models, resolutions, durations, aspect ratios, and optional keyframes for advanced video control.

Configuration and environment

Configure the MCP client to reference the Luma MCP server. The example setup places the server under the key luma with a stdio command that launches the server project and passes the API key via environment variables.

Troubleshooting and notes

If you encounter issues, verify that your LUMA_API_KEY is correct and that the server path is accessible. Review startup logs to confirm that the MCP server starts correctly and that the runtime runner can locate the project files. If problems persist, ensure the Dream Machine API v1 endpoint is reachable and that your account has the necessary permissions and credits.

Security and best practices

Keep your Luma API key secure. Do not commit the key into code or version control. Use environment variables to pass sensitive credentials and restrict access to the MCP server to trusted clients.

Available tools

ping

Check if the Luma API is running; takes no parameters and returns a simple status.

create_generation

Create a new video generation from a text prompt with options for model, resolution, duration, aspect ratio, looping, and advanced keyframes.

get_generation

Retrieve the status and details of a specific generation by its ID, including state, failure reason, and resulting video URL if completed.

list_generations

List existing generations with optional pagination (limit and offset) to manage history and tracking.

delete_generation

Delete a specific generation by its ID to clean up resources.

upscale_generation

Upscale a completed generation to a higher resolution, with a constraint that it can only happen once per generation and the target resolution must be higher.

add_audio

Attach AI-generated audio to a video generation using a provided prompt, with optional negative prompt and a callback URL.

generate_image

Create an image from a text prompt, with optional reference images and style references to guide the result.

get_credits

Return the current user's available credits in USD cents.

get_camera_motions

Return the list of supported camera motion options for tooling and animation