Odysseus on Windows: setup with Ollama

This tutorial covers native Windows installation of PewDiePie's Odysseus and connects it to Ollama for a local LLM chat session. The title specifies Windows 10/11. The presenter briefly notes Docker and other platform instructions, then follows the Windows route with Git and Python 3.11 or later.

The setup starts by cloning the repository and running its PowerShell installer. The script checks Python, creates a virtual environment, installs packages, and launches first-time setup. After creating an administrator username and password, the presenter opens localhost:7000 and signs into the chat interface.

The model settings support local endpoints and API providers. The demonstration uses automatic Ollama detection to find an existing Gemma model, adds the endpoint, and receives a chat response. LM Studio also appears as a local connection option. The presenter describes a cookbook feature that scans hardware and suggests models to download, though the demonstrated connection uses an existing Ollama installation.

A brief tour covers deep research controls, including research rounds, search engine selection, and the model endpoint. The final section shows how to restart the app by activating its virtual environment and running the documented Python command. It also shows an execution-policy error during activation and the presenter's PowerShell workaround.