Jan
Jan is an open-source desktop application that serves as a ChatGPT alternative, allowing users to download and run large language models locally with full control and privacy. It also supports cloud model integration and provides an OpenAI-compatible API for developers. It is designed for users who want to run AI models on their own hardware or prefer a privacy-focused AI assistant.
✨ Key features
- Run local LLMs from HuggingFace (Llama, Gemma, Qwen, GPT-oss).
- Connect to cloud models: OpenAI, Anthropic, Mistral, Groq, MiniMax.
- Create custom AI assistants for specialized tasks.
- OpenAI-compatible local server at localhost:1337.
- Model Context Protocol (MCP) integration for agentic capabilities.
- Privacy-first: everything runs locally when desired.
🎯 Use cases
- Run LLMs offline on personal computer for privacy.
- Develop applications using OpenAI-compatible API locally.
- Create custom assistants for specific workflows.
- Integrate with MCP for agentic AI capabilities.
- Compare local and cloud model performance.
📦 Installation
🧰 Requirements: Desktop app available for Windows, macOS, Linux; minimum specs: macOS 13.6+, Windows 10+ with GPU support, Linux; RAM varies by model size (8GB for 3B, 16GB for 7B, 32GB for 13B).
Download the appropriate installer from jan.ai or GitHub Releases:
- Windows: jan.exe
- macOS: jan.dmg
- Linux (deb): jan.deb
- Linux (AppImage): jan.AppImage
- Linux (Arm64): see How-to
For building from source, prerequisites: Node.js ≥ 20.0.0, Yarn ≥ 4.5.3, Make ≥ 3.81, Rust (for Tauri), and on macOS Apple Silicon MetalToolchain.
git clone https://github.com/janhq/jan
cd jan
make dev
This installs dependencies, builds core components, and launches the app. Manual commands: yarn install, yarn build, yarn dev.
🚀 Usage
After installation, launch Jan. Download a model from HuggingFace or connect to a cloud provider. Use the OpenAI-compatible API at localhost:1337 for other applications. For development, run make dev to start the app in development mode.
⚠️ Good to know
System requirements vary by model size; GPU acceleration is recommended for Windows and Linux; building from source on Windows requires specific tools and may hit path length limits.
❓ FAQ
What are the system requirements for Jan?
Minimum specs: macOS 13.6+ (8GB RAM for 3B models, 16GB for 7B, 32GB for 13B), Windows 10+ with GPU support for NVIDIA/AMD/Intel Arc, Linux most distributions work with GPU acceleration available.
Can I use Jan with cloud models?
Yes, Jan supports connecting to GPT models via OpenAI, Claude via Anthropic, Mistral, Groq, MiniMax, and others.
How do I build Jan from source?
Clone the repository, ensure prerequisites (Node.js, Yarn, Make, Rust) are installed, then run make dev which handles everything. Manual commands are yarn install, yarn build, yarn dev.
Does Jan provide an API for developers?
Yes, Jan runs an OpenAI-compatible local server at localhost:1337 that other applications can use.
📊 Repository
🤖 Overview, features, install steps and FAQ were generated from the project's README on Sep 4, 2026. Always check the original source before running commands.