Overview

Jan is an open-source desktop application that runs large language models locally on your computer. Built with the user experience of commercial AI assistants in mind, it offers a clean chat, a model hub, and full offline operation — no account, no cloud, and no data leaving your machine. Jan aims to be the free, open alternative to closed desktop AI apps, with an extensible architecture and a growing ecosystem of plugins and integrations.

Key Features

• 100% Local & Private: Everything runs on your hardware; works completely offline
• Model Hub: Browse and download thousands of models, including GGUF formats, from inside the app
• Polished Chat UI: A modern assistant interface with streaming responses and conversation history
• OpenAI-Compatible API: Expose a local API server for integrating with other applications
• Cross-Platform: Native desktop apps for Windows, macOS, and Linux
• Extensible: Plugin architecture and open codebase allow customisation and contributions
• Multi-Modal Support: Run vision and other model architectures alongside text models

Why Use It

Jan fills the same role as closed local AI apps but stays fully open source — the desktop UI and engine are on GitHub under an Apache 2.0 licence. If you want the convenience of a local assistant app without proprietary lock-in, Jan is a direct, free alternative. Its privacy story is strong: models run on your machine, so nothing is uploaded, and there is no commercial service dependency for basic use.

Use Cases

• Offline assistant: Use a capable model for writing, coding, and Q&A with no internet
• Private data work: Process sensitive documents without sending them to any server
• Local prototyping: Back your scripts and tools with a local OpenAI-compatible endpoint
• Model testing: Try multiple open-weight models on your own hardware

Platform

Windows, macOS, Linux

Licence

Apache License 2.0 (open source)

Website

jan.ai · github.com/menloresearch/jan