Jan: Run Local AI on Your Desktop with llama.cpp

Overview

Jan is a Tauri desktop app built on llama.cpp that makes open-source AI models available in a ChatGPT-style interface. It is aimed at people who want to download and run models locally, with control over their data and the option to work offline.

The app supports Windows, macOS, and Linux. Users can also connect cloud AI services, create custom assistants, and expose a local OpenAI-compatible API, so Jan can serve as both a desktop chat app and a local AI endpoint for other tools.

Key Features

• Download and run models such as Llama, Gemma, Qwen, and GPT-oss through Hugging Face.
• Run models locally for privacy and offline use.
• Connect to cloud services including OpenAI, Anthropic, Mistral, Groq, and MiniMax.
• Create custom assistants for specific tasks.
• Provide an OpenAI-compatible API on localhost:1337.
• Integrate with the Model Context Protocol for agentic workflows.
• Install on Windows, macOS, or Linux.

Use Cases

• Individuals who want to chat with local models without sending prompts to a cloud service.
• Developers who need a local OpenAI-compatible endpoint to test applications against models on their own machine.
• Teams that want a desktop interface for trying different downloaded models and custom assistants.
• Users who want one app for both local models and supported cloud AI services.

Getting Started

Download Jan for your platform from jan.ai or GitHub Releases. For build instructions and platform-specific guidance, see the README.

License in Plain Words

No standard license was detected. Without a license, the default copyright rules apply: you may read the code, but not reuse it. Check the repository or ask the authors.

Alternatives

• OGAD: OGAD focuses on a local-first studio and gateway for text, vision, image, and voice models, while Jan centers on desktop chat with local or cloud providers.
• page-assist: Page Assist puts local AI chat in a browser sidebar and can answer questions about the page you are viewing; Jan is a standalone desktop chat app.
• textgen: TextGen offers a desktop and web interface with vision and tools for locally run models, while Jan also connects to cloud AI providers.

Considerations

• Running local models depends on your hardware and the model size. The README gives memory guidance for several model sizes, and notes GPU support for Windows and available GPU acceleration on Linux.
• Building from source requires Node.js, Yarn, Make, and Rust, with additional platform-specific tooling for some builds.
• The project is actively maintained, with recent commits and releases and a large contributor base; its measured contributor activity is not concentrated in one person.
• Confirm the applicable license before adopting or redistributing the project: the README says Apache 2.0, but repository metadata reports NOASSERTION.