Canon MCP

What is Canon MCP?

Canon MCP is a server that lets you control Canon cameras supporting CCAPI (Camera Control API) using natural language commands processed by Large Language Models. It replaces traditional camera interfaces with conversational interactions, ideal for developers and photographers wanting AI-assisted remote camera control.

How to use Canon MCP?

Clone the repository, run npm install && npm build && npm build-container, then add the server to your MCP config file with the command node /ABSOLUTE/PATH/TO/REPO/build/mcp.js. The server exposes tools like connect-camera, take-photo, set-iso-setting, and start-camera-livestream that can be invoked via any MCP client.

Key features of Canon MCP

• Connect to Canon cameras over CCAPI
• Take photos and start interval shooting
• Adjust shooting settings (ISO, shutter speed, aperture)
• Control autofocus, white balance, and color temperature
• Get live view images and start a browser livestream
• Retrieve camera status (battery, storage, temperature)

Use cases of Canon MCP

• Remotely control a Canon camera from an AI assistant
• Automate time‑lapse photography with interval photo commands
• Change shooting modes and settings without touching the camera
• Monitor camera status (battery, storage, lens info) via chat

FAQ from Canon MCP

What cameras are supported?

Any Canon camera that has CCAPI enabled via a Canon‑provided firmware update. You must activate CCAPI on your camera first.

What are the system requirements?

You need Node.js to build and run the server, and a Canon camera with CCAPI firmware. The camera must be reachable over the network.

How do I connect to a camera?

Use the connect-camera tool with the camera’s IP address and port (e.g., http://10.0.0.241:8080).

Is this project stable?

The project is under active development; features and documentation may be incomplete or change.

Where are photos saved?

Photos taken via the server are saved to the output directory specified in the configuration (the get-last-photo tool stores the photo locally).