AprilCam

AprilCam gives AI agents and robot programs a shared, real-time picture of a robotics playfield — AprilTag/ArUco positions, orientation, velocity, and homography — without each consumer running its own vision pipeline.

Architecture

A single long-running daemon (aprilcamd) owns the cameras, runs detection, and serves per-frame data. Two front ends sit on top of it:

• MCP server (aprilcam mcp) — tools for AI agents to open cameras, create playfields, query tags, and run image-processing operations over the Model Context Protocol.
• Direct gRPC client (DaemonControl) — a low-latency Python API for robot control loops that read tags and draw live overlays at 5–50 Hz.

Both talk to the daemon over the same gRPC service and protobuf stream sockets.

Where to start

• Robot Direct API — the Python client for high-frequency tag reads and live overlay drawing in a control loop.
• Daemon Wire Protocol — the gRPC service and protobuf stream framing, for building a client in another language.
• Tag Detection Under Variable Lighting — why tags drop out under glare and the preprocessing pipeline that recovers them.

Install & run

pipx install 'aprilcam[daemon]'   # full server stack
aprilcam mcp                       # run the MCP server (stdio)
aprilcam daemon start              # or run the daemon directly

See the repository README for install tiers, the full MCP tool reference, and CLI commands.