janhq/cortex.cpp

Cortex is an open-source local AI API platform designed to empower users to run a diverse range of AI models, including vision, speech, language, tabular, and action models, directly on their local hardware. It offers robust multi-engine support, initially integrating with `llama.cpp`, and features automatic hardware optimization for various GPUs from NVIDIA, AMD, and Intel. By providing an OpenAI-compatible API, Cortex aims to serve as a flexible and powerful "brain for robots" that operates independently of cloud services. This approach offers a compelling solution for local AI development, experimentation, and deployment, enabling greater privacy, control, and efficiency for AI applications.