LM Studio
LM Studio is a free desktop application for running large language models locally without cloud connectivity or subscriptions. Features polished GUI with model browser for HuggingFace search/download, visual performance comparisons, immediate chat interface, user-friendly parameter sliders, and automatic hardware detection. Intelligent memory management with Vulkan offloading for integrated Intel/AMD GPUs. Excellent Apple Silicon optimization with Metal acceleration. OpenAI-compatible local API server with streaming and embeddings support. Uses llama.cpp inference engine. Perfect for beginners and users preferring graphical interfaces. Available for macOS, Windows, and Linux.