🚧 Cortex is under construction.


🚧 Cortex is under construction.

Cortex Cover Image

Cortex is an OpenAI compatible (opens in a new tab), local AI server that developers can use to build LLM apps. It can be used as a standalone server, or imported as a library.

Cortex currently supports two inference engines:

  • Llama.cpp
  • TensorRT-LLM

Real-world Use: Cortex powers Jan, our local ChatGPT-alternative.

Cortex has been battle-tested through 900k downloads, and handles a variety of hardware and software edge cases.


Cortex's roadmap is to implement an OpenAI-equivalent API (opens in a new tab) using a fully open source stack. Our goal is to make switching to open source AI as easy as possible for developers.


Cortex's architecture features C++ inference core, with higher-order features handled in Typescript.

Our long-term direction is to (eventually) move towards being a full C++ library to enable embedded and robotics use cases.