How to use Local WebAssembly LLM Chat

• 1Select a local model architecture from the WebGPU Engine Config panel.
• 2Click 'Download Weights & Load Engine' to simulate moving the model into your browser's VRAM.
• 3Upload text or markdown files into the Local Vector Store to enable local RAG (Retrieval-Augmented Generation).
• 4Ask a question in the chat box and watch the model stream a response at over 40 tokens/sec, totally offline!