RCLI: On-Device Voice AI for macOS with Local RAG and Vision Capabilities

Summary

RCLI is an on-device voice AI application for macOS that provides a complete AI pipeline running locally on Apple Silicon. It features speech-to-text, large language models, text-to-speech, and vision language models all operating natively without cloud dependency. The tool offers 40 macOS voice actions, local RAG (Retrieval-Augmented Generation) for document querying, on-device vision capabilities for camera and screen analysis, and achieves sub-200ms end-to-end latency. Powered by MetalRT, a proprietary GPU inference engine optimized for Apple Silicon, RCLI enables natural voice conversations with the Mac while maintaining privacy and speed through local processing.