Enfra is open-source infrastructure for running frontier LLMs and VLMs on edge devices — mobile, robots, autonomous vehicles, and beyond.
Optimized for ARM, RISC-V, and mobile GPUs. Deploy anywhere — from embedded devices to wearables.
INT4, INT8, Q4_K_M, and ONNX export out of the box. Compile to your target with a single command.
GGUF, ONNX, CoreML, and TFLite. Built for mobile SoCs, NPUs, and edge inference with minimal memory footprint.