Claude Glass (Or Black Mirror)

Build a real-time vision processing agent that uses a Claude-like visual reasoning model to act as a 'Claude Glass' for developers. This tool would take a live video feed, apply stylistic or analytical filters via LLM vision analysis, and output augmented data to a browser or HUD overlay.

Suggested repo: glass-vision

"Turn your webcam into a context-aware reasoning engine that sees the world like an LLM."

Estimated effort: 40h