How to install PewDiePie's Odysseus AI on your own computer

Odysseus is the AI workspace PewDiePie built for himself and then gave away for free. It does chat, deep research, email, notes and documents in one app, and it runs on your own computer instead of a company's servers. His own AI model, Ajax, will run inside it when it's released.

This guide gets Odysseus running in three steps. You can let your AI assistant do the typing, or follow the commands yourself. It takes 15 to 30 minutes, most of it waiting for downloads.

Before you start

• A computer running Windows, macOS or Linux. Odysseus itself is light. Running AI models on your own computer is the heavy part, so a recent graphics card or an Apple Silicon Mac with 16 GB of memory or more helps. On a weaker computer you can connect it to an online model instead.
• Git, which copies the project to your computer. It's free from git-scm.com.
• Python 3.11 or newer (python.org) if you use the Windows or Mac route, or Docker Desktop (docker.com) if you use the Docker route.

Only download from the official project. Odysseus lives at github.com/odysseus-dev/odysseus. PewDiePie's old link, github.com/pewdiepie-archdaemon/odysseus, redirects there. There's no official .exe or app download, so if a site offers an "Odysseus setup file", it isn't the real one.

Step 1: Find it on GitHub

Go to github.com/odysseus-dev/odysseus, or search "odysseus-dev" on GitHub. It's the one with tens of thousands of stars, owned by odysseus-dev. Everything you need is in that project.

Step 2: Install it

The easy way: let your AI assistant do it

If you use an AI coding assistant that can run commands on your computer, such as Claude Code, Cursor or GitHub Copilot, paste this in:

Install Odysseus for me. It's a free, self-hosted AI workspace.

Only use the official repository: https://github.com/odysseus-dev/odysseus
Read its README and website/setup.md first, then:
1. Check I have Git and either Docker Desktop or Python 3.11+. Tell me plainly if anything is missing and how to get it.
2. Clone the repository into a folder in my home directory.
3. Install and start it the way setup.md recommends for my operating system.
4. Find the temporary admin password it prints on first start and show it to me.
5. Tell me the exact address to open in my browser.

Keep the login switched on and don't open it to my network or the internet. I'm not a developer, so explain each step in one simple sentence.

Your assistant will tell you what it's doing at each step. When it's done, it gives you an address and a password. Skip to step 3.

Do it yourself: Windows

Open PowerShell and run these three lines one at a time:

git clone https://github.com/odysseus-dev/odysseus.git
cd odysseus
powershell -ExecutionPolicy Bypass -File .\launch-windows.ps1

The launcher sets everything up and starts Odysseus. You can run it again any time to start it.

Do it yourself: Mac

Open Terminal and run:

git clone https://github.com/odysseus-dev/odysseus.git
cd odysseus
./start-macos.sh

On a Mac this route is better than Docker, because Docker can't use the Mac's graphics chip.

Do it yourself: Docker (any computer)

With Docker Desktop running, open a terminal and run:

git clone https://github.com/odysseus-dev/odysseus.git
cd odysseus
cp .env.example .env
docker compose up -d --build

The first time takes a while because it downloads everything it needs. To see the admin password, run:

docker compose logs odysseus

Step 3: Open it in your browser

The address depends on how you installed it:

• Windows or Linux launcher: http://localhost:7000
• Mac start script: http://localhost:7860 (macOS uses port 7000 for AirPlay)
• Docker: http://localhost:7011

Log in as admin with the temporary password Odysseus printed when it started, then change the password in Settings.

Now pick an AI model. Open Cookbook: it checks your graphics card and memory, scores the models that will actually run on your computer, then downloads and starts the one you choose. If you already use Ollama, you can add it in Settings instead, at http://localhost:11434/v1, and use the models you already have.

That's it. You have a private AI workspace that runs on your own computer.

Keep it safe

• Keep the login switched on. Odysseus's agents can run commands on your computer, so don't turn off the password.
• Keep it on your own computer. By default it only listens to your own machine. Don't open it to your network or the internet.
• Update it now and then. Odysseus changes almost every day. In the Odysseus folder, run git pull and start it again.

What about Ajax?

Ajax is the small AI model PewDiePie trained to run inside Odysseus. As of 20 October 2026, data.pewdiepie.com still says "Coming soon". When it's out, it will show up as a model you can run inside Odysseus. Until then, Cookbook can run similar small open models, including Qwen models like the one Ajax is built on. Read our Ajax article for the full story.