Llamafile

Llamafile is a Mozilla project that lets you run LLMs with a single executable file — no installation, no dependencies, no setup. It bundles model weights and the inference engine into one portable file that runs on macOS, Linux, and Windows. Claims the fastest prompt processing speed on consumer hardware. Simply download the file, make it executable, and run. Perfect for distributing AI models to non-technical users or air-gapped environments.