Windows install

How to install Odysseus AI on Windows with WSL2

For most Windows users, WSL2 plus Docker Desktop is the cleanest way to run Odysseus. It avoids many Python and path issues while keeping local model serving available through Ollama.

Requirements

Windows 11 or a current Windows 10 build with WSL2 enabled.
Docker Desktop with WSL integration enabled for your Linux distro.
At least 16 GB system memory for comfortable testing.
Optional NVIDIA GPU with current drivers if you want faster local inference.

Recommended flow

Install WSL2 first, then Docker Desktop, then verify Docker works from inside the WSL terminal. After that, install or connect Ollama, pull a small model, and only then launch Odysseus. This order makes debugging easier because each layer can be tested separately.

wsl --install
wsl --status
docker version
ollama list

NVIDIA GPU notes

If Odysseus or Ollama cannot see your GPU, the problem is usually outside Odysseus. Confirm Windows NVIDIA drivers, WSL GPU visibility, Docker Desktop WSL integration, and NVIDIA container runtime support before changing Odysseus settings.

Security note: Do not expose your Odysseus instance directly to the public internet until authentication, reverse proxy headers, TLS, and private network rules are reviewed.

First-run checks

Open the local Odysseus URL in a browser and confirm the login page appears.
Check the generated admin password or setup output before closing the terminal.
Verify a model appears in the model selector after connecting Ollama.
Save a copy of your environment variables and volume paths.

Next: read the model setup guide or jump to common Odysseus fixes.