Jan is an open-source desktop application for running and using AI models locally, with an experience similar to a private ChatGPT-style assistant. It lets users download compatible models, chat with them on their own computer, and connect to external model providers when local hardware is not sufficient. The project is designed around user control, offline capability, and an open ecosystem rather than locking conversations to one commercial AI service.

Jan includes a desktop interface, local model management, an OpenAI-compatible API server, and extension points for integrating AI into other applications. It is particularly useful for users who want private local inference, developers testing multiple open models, and teams that need a familiar chat interface without sending every conversation to a hosted provider. Performance depends on the selected model and available hardware, but the platform lowers the barrier to experimenting with local LLMs.