Jan (Jan.ai) is a free and open-source tool for running large models locally and having AI chat conversations. It helps users install, deploy, run, and use open-source ChatGPT alternative large models on their local computers (Windows, Mac, Linux), such as LLaMa, Mistral, Phi-2, and more than 20 other models. It also supports entering your own OpenAI API Key to run GPT. Compared with those previously introduced by the AI tool collectionOllama, this tool provides a chat UI and an API server, making it suitable for developers, researchers, or AI enthusiasts to experience open-source large models locally.
Jan's main features
- Run open-source large models locally: Supports more than 20 models including LlaMa, Mistral, Phi-2, DeepSeek, Yi, etc. They can be imported manually or downloaded from the model library
- Clean and easy-to-use chat interface: The software interface is fresh, simple, intuitive, and easy to use, allowing you to quickly chat with open-source large models directly on your local machine
- Supports multiple platforms: Jan supports running on operating systems such as Windows, Mac (Intel, M1/M2/M3), and Linux, and a mobile app will also be released in the future.
- Built-in API server: Compatible with the OpenAI API, allowing you to get model information, download, start, stop models, and chat through the API, etc.
How to use Jan
- Visit Jan's official website (jan.ai), select the version for your computer's operating system, and click Download to download it
- Then install and open the software, and click Download your first model in the lower-left corner of the software interface
- Enter the model library Hub interface, select the model you are interested in and download it. After the download is complete, click Use to use that model
- In the chat interface, enter your description and you can chat with the model you selected
Jan is still under continuous development, and will later support features such as a mobile app, creating AI assistants, inference engines, and plugin extensions.
