Cameracontroller

What is Cameracontroller?

Cameracontroller is a multi‑protocol camera control service built on Node.js/TypeScript that follows the Model Context Protocol (MCP). It enables AI assistants to enumerate cameras, capture photos, record videos, and start live video streams on Windows and Linux.

How to use Cameracontroller?

Install via npm install, copy .env.sample to .env, configure environment variables, build with npm run build, then start with npm start (or npm run start:http for HTTP/SSE mode). For testing, use npm link and connect with the MCP Inspector using STDIO transport.

Key features of Cameracontroller

• Camera enumeration, photo capture, and video recording
• Video streaming with streamUrl support
• MCP tools: listCameras, takePhoto, startVideo, stopVideo
• Multiple transport modes: stdio, HTTP, SSE
• Supports Windows (dshow) and Linux (v4l2)
• Concurrent multi-instance recording and streaming

Use cases of Cameracontroller

• Take a photo from a specific camera and save it.
• Start a video recording to a file with an optional duration.
• Begin a live video stream accessible via a temporary URL.
• Stop an ongoing recording or stream by session ID.
• Enumerate all available cameras before selecting one.

FAQ from Cameracontroller

What transport modes does Cameracontroller support?

It supports stdio, HTTP, and SSE (Server-Sent Events) transports.

What dependencies are required to run Cameracontroller?

Node.js 16+ and ffmpeg (in PATH or provided by ffmpeg-static) are required.

How can I test the MCP tools?

Use the MCP Inspector (npx modelcontextprotocol/inspector) with STDIO transport and the command npx mcp-camera-server, or connect via HTTP/SSE.

Why does the streamUrl expire or work only once?

By design, each streamUrl can be consumed only once and has a configurable timeout (default 300 seconds). Call startVideo again to get a new streamUrl.

Can I record multiple videos at the same time?

Yes, Cameracontroller supports concurrent multi-instance recording and streaming.