JanJan
Docs
Jan Desktop
Install 👋 Jan
Linux

Linux Installation

Instructions for installing Jan on Linux.

Install Jan

Flatpak is the recommended way to install Jan on Linux. It runs on any distribution and updates automatically through Flathub. Use the .deb or AppImage only if you can't use Flatpak.

Install from Flathub (opens in a new tab):


flatpak install flathub ai.jan.Jan

Then launch it:


flatpak run ai.jan.Jan

If you don't have the Flathub remote set up yet:


flatpak remote-add --if-not-exists flathub https://flathub.org/repo/flathub.flatpakrepo

Updating Jan

How Jan updates depends on which package you installed:

Flatpak updates through Flathub along with the rest of your system. The in-app updater is disabled — your package manager handles it:


flatpak update ai.jan.Jan

Compatibility

  • OS: Debian/Ubuntu, Fedora/RHEL, Arch, openSUSE
  • CPU: AVX2 required — Intel Haswell (2013+), AMD Excavator (2015+)
  • Memory: 8GB minimum (16GB recommended)
  • GPU: 6GB VRAM minimum
  • Storage: 10GB free space minimum

Jan Data Folder

Jan stores all your data locally — downloaded models, conversation threads, settings, and logs generated from using the app. Nothing is sent to the cloud.

Default location (resolves via XDG_DATA_HOME):


~/.local/share/Jan/data

The path to Jan's per-user configuration file (which records the data folder location) lives at ~/.config/Jan/.

See Jan Data Folder for details.

GPU Acceleration

Configuration for GPU support:

Step 1: Verify Hardware & Install Dependencies

1.1. Check GPU Detection


lspci | grep -i nvidia

1.2. Install Required components

NVIDIA Driver:

  1. Install the NVIDIA Driver (opens in a new tab), ideally via your package manager.
  2. Verify:

nvidia-smi

CUDA Toolkit:

  1. Install the CUDA toolkit (opens in a new tab), ideally from your package manager (11.7+)
  2. Verify:

nvcc --version

Additional Requirements:


sudo apt update
sudo apt install gcc-11 g++-11 cpp-11
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64

Documentation (opens in a new tab)

Step 2: Enable GPU Acceleration

  1. Navigate to Settings () > Model Providers > Llama.cpp
  2. Select an appropriate backend in Backend. Details in our llama.cpp guide.

CUDA offers better performance than Vulkan.

Uninstall Jan

Removal commands:


flatpak uninstall ai.jan.Jan
rm -rf ~/.local/share/Jan # data folder (models, threads, logs)
rm -rf ~/.config/Jan # app configuration

⚠️

These commands permanently delete application data.

WindowsJan-Code-4B