Large Language Models

Framework for building LLM-powered applications using composable primitives.

Run LLaMA-class models efficiently on consumer hardware.

Simple local LLM runtime with a clean developer experience.

High-throughput, memory-efficient LLM serving engine.