coreai-onnx#

coreai-onnx converts ONNX models into Apple’s Core AI .aimodel format for on-device inference on Apple silicon. It supports 143 built-in ONNX operator lowerings, graph optimization, and numerical verification against ONNX Runtime.

Convert models exported from PyTorch and other ONNX-compatible frameworks through the Python API, command-line interface (CLI), or Model Context Protocol (MCP) server. Conversion runs on macOS, Linux, and Windows; executing .aimodel bundles requires macOS 27+ / iOS 27+ with the Core AI framework.

Get started · Supported ONNX operators · MCP server · GitHub · PyPI