Skip to content
fedi.software

Jan for Windows: questions and answers

0.8.4 · Menlo Research · Page updated

Open source · Apache-2.0 Windows

Jan is an open-source desktop chat app for Windows that runs language models offline and can also connect to cloud APIs. Version 0.8.4, Apache-2.0 licence.

Download Official download link

Rate Jan

Thank you! Your rating is saved; you can change it. Could not save the rating. Please try again later. Too many ratings from your network today. Please try again tomorrow. The average appears after 5 ratings.

About Jan

Jan 0.8.4 for Windows is an open-source chat app in the style of ChatGPT that runs language models directly on your computer. It is developed by Menlo Research and published under the Apache-2.0 licence. The Windows installer is small, about 55 MB, and the download button here uses the official app.jan.ai link to the latest release, which is the same file as the asset on GitHub.

What is Jan?

Jan gives you a familiar chat interface — conversations on the left, the thread in the middle, model settings on the right — but the model answering you can live on your own hard drive. Offline models run through llama.cpp, the engine behind many local AI tools. When you want a cloud model instead, you add an API key for a provider and switch in the same window, so local and remote models sit side by side. Jan can also act as a local server with an OpenAI-compatible API for other apps.

Does Jan really work offline?

Yes, for local models. Once a model is downloaded, chatting with it needs no internet connection and your prompts do not leave the computer. Cloud models naturally need a connection and send your messages to the provider you chose, under that provider's terms.

What hardware do I need?

  • Windows 10 or later; Jan is also available for macOS and Linux.
  • About 8 GB of RAM is enough for small models of up to 3B parameters; larger models need more.
  • A GPU is optional: NVIDIA, AMD and Intel Arc cards are supported through Vulkan.

The system requirements in the repository are kept up to date by the developers.

How do I set it up?

  1. Download and run the installer, Jan_0.8.4_x64-setup.exe.
  2. Open the model hub inside the app and choose a model that fits your memory.
  3. Wait for the download, then start a new chat and select that model.
  4. Optionally add API keys for cloud providers in the settings.

Is Jan free?

The app is free and open source, with its code on GitHub. Local models are free to run, apart from their licence terms. Cloud models are billed by their providers if you connect them.

How recent is version 0.8.4?

Version 0.8.4 was released on 23 July 2026. The list of releases, with notes for each one, is on the project's releases page.

How does Jan compare with similar apps?

Jan and LM Studio look alike at first glance, but Jan is open source and treats cloud APIs as equal citizens next to local models. Ollama is more of an engine with a terminal and an API than a chat app. GPT4All centres on chatting with local documents, and AnythingLLM adds workspaces and agents. If you want one tidy window where a local model and, say, a hosted model can be swapped mid-conversation, Jan is built for exactly that.

Advantages and disadvantages

  • Advantage: open source under the Apache licence.
  • Advantage: local and cloud models in the same interface.
  • Advantage: a small installer.
  • Disadvantage: models must be downloaded separately and can be large.
  • Disadvantage: speed depends heavily on your hardware.