zoom_in Click to enlarge

Jan

▲ ▼

language

description Jan Overview

Jan is a self-hosted open source AI assistant designed for local use. It provides a GPT-alternative experience running on desktop computers.

This allows developers and users to utilize an API for tasks like code completion and assistance within tools such as JetBrains IDEs. Jan offers a private and customizable solution compared to cloud based services.

help Jan FAQ

Does Jan run AI models locally on Windows, macOS, and Linux?

Yes. Jan provides desktop applications for Windows, macOS, and Linux, and its local models can run entirely on the user's computer without an account or API key. Cloud providers are optional rather than required. [Jan](https://www.jan.ai/docs/desktop/quickstart)

Which local model format does Jan use?

Jan manages local models through llama.cpp, with GGUF as the main model format. Performance depends on the computer's RAM, processor, and graphics hardware, so Jan shows hardware-fit warnings in its model hub. [Jan](https://jan.ai/docs/manage-models)

Can Jan connect to a JetBrains IDE for code completion?

Jan does not need a special JetBrains plugin to expose a model, because it can serve an OpenAI-compatible local API. The current Jan CLI documentation uses localhost:6767/v1, and an IDE plugin that supports a custom OpenAI endpoint can be configured to use it. [Jan CLI](https://www.jan.ai/docs/desktop/cli)

Can Jan use cloud models as well as local models?

Yes. The desktop app can run local llama.cpp models or connect to providers such as OpenAI, Anthropic, Groq, and Mistral through compatible API settings. Local use is private and offline, while cloud use requires the provider's API key. [Jan](https://www.jan.ai/docs/desktop/quickstart)

swap_horiz

Looking for Jan alternatives? Compare top competitors ranked & scored

arrow_forward