Cortex

Cortex  turns any computer into a thinking machine.

• Simplifies running of models
• Replaces the OpenAI API with a local API
• Runs on wide variety of hardware (from Raspberry Pis to servers)

Cortex powers Jan’s local AI capabilities, and runs quietly in the background.

Model Hub

Cortex has a Model Library , where we have pre-compiled and optimized models.

Roadmap

Our current roadmap is fully public on GitHub .

• Support vLLM and continuous batching
• Support Vision, speech, language models
• Support additional hardware types
• Scale from single-node edge devices, to multi-GPU datacenter environments
• Tools for managing Cortex deployments

Open-Source

Cortex builds on llama.cpp , vLLM  and other open source projects.

We have hired from the llama.cpp contributor communities, to allow open-source developers to focus on this full-time.