PewDiePie Launches Odysseus, a Free AI Interface That Runs on Your Own PC
YouTuber PewDiePie has released Odysseus, an open-source platform for running AI language models on your own computer. Your chats, files and settings stay local, and he says there's no telemetry. He pitches the project as an alternative to paying for ChatGPT.

What's going on
Felix Kjellberg, better known online as PewDiePie and one of the biggest YouTubers in the world, has released a software project of his own: Odysseus. It's a free, open-source platform for running AI language models on your own computer instead of sending your questions to cloud services like ChatGPT or Claude. He announced it in a YouTube video with the tongue-in-cheek title "MY trillion $Dollar Project is finally OUT!" According to the reports, the code is publicly available on GitHub. Anyone can download it, modify it and run it themselves.
A note on timing: according to The Business Standard (tbsnews.net), the video went up on May 31. Swedish site Feber only picked up the story in November. However, our main source from Feber doesn't contain its own article on Odysseus under that headline, so this piece relies on reporting from Mezha and The Business Standard.
Why PewDiePie is building this
PewDiePie has spent the past few months experimenting publicly with Linux and local language models. According to The Business Standard, he spent about a year documenting on his channel how he builds his own machines to run open-source language models. Mezha reports that he explains his motivation like this: he dislikes the business model of popular AI services, which learn a great deal about their users from queries, uploaded documents and chat histories.
Odysseus is meant to offer the convenience of well-known AI services without relying on someone else's servers or handing personal data to big tech companies. The official description reads:
"Odysseus is a self-hosted interface for talking to language models – chat, autonomous agents, tools, model serving, email, research and more. Local-first, privacy-first, no telemetry. Just you and your models."
"Self-hosted" means the software runs on your own device or server, not with a provider. "Telemetry" refers to usage data that programs send back to the developer in the background. Odysseus says it doesn't collect any at all.
What Odysseus can do
Odysseus isn't a language model itself. It's an interface, a kind of cockpit for controlling different models. For the actual number crunching, it relies on well-known programs that run models locally:
- Ollama: a popular, beginner-friendly tool that downloads and starts models with a single command.
- llama.cpp: lean software that runs models efficiently even on ordinary CPUs and graphics cards.
- vLLM: software built for speed, aimed more at powerful GPUs and servers.
If you also want to use large cloud models, Mezha says you can connect OpenAI or OpenRouter. OpenRouter is a service that offers many different AI models through a single shared interface.
According to the reports, features include:
| Feature | What it does |
|---|---|
| Chat | Conversations with local and cloud models |
| Agent mode | The model uses tools to carry out several steps on its own, e.g. a research task |
| Memory | The AI remembers information across conversations |
| Files | Upload documents and work with them |
| Automatic model selection | Odysseus suggests models that suit your PC's hardware |
| Model comparison | Compare answers from several models side by side |
| Email and research | Built in, according to the official description |
In the launch video, The Business Standard reports, PewDiePie showed the platform handling research tasks, managing conversations and serving as a private AI assistant without depending entirely on third parties. Automatic model selection should be especially handy for beginners. Figuring out which model will run well on which machine is often the biggest hurdle with local AI.
Built with AI, positioned against Big Tech
According to Mezha, PewDiePie openly admits that much of the project was built with AI help. The repository names Claude, DeepSeek and Qwen, among others, as models that helped with development. That's become a familiar pattern: with AI coding assistants, even individuals without a big dev team can build substantial software.
The project's tone is combative. The Odysseus website says, according to The Business Standard: "No sales team, no demo request, no Trojan horse." In the video, PewDiePie said:
"The war against Big Tech has only just begun."
At the same time, Mezha reports, he stresses that his main goal isn't to compete with ChatGPT or Claude. He wants users to have more control over their own data and to be able to run AI on their own hardware.
Watch out for copycats
Wherever a famous name shows up, opportunists follow. Crypto articles are already circulating under the headline "Pewdiepie Odysseus AI," supposedly offering price predictions for memecoins like PEPE. They have nothing to do with the actual project: Odysseus is a chat and agent interface, not a crypto forecasting tool. Those pieces are speculation and advertising. If you want to try Odysseus, get it only from the official GitHub repository or the official project site.
Our take
Technically, Odysseus doesn't reinvent the wheel. Self-hosted AI interfaces have been around for a while, including Open WebUI, LM Studio and AnythingLLM. They offer similar features such as chat, documents and agents. What makes Odysseus stand out is mainly its creator's reach. PewDiePie speaks to a huge, young audience that has barely touched local AI so far. If even a fraction of them start running language models on their own machines, that would give AI privacy a real boost.
Some questions remain open. The sources include no independent tests of stability, security or usability. That matters for a project largely written by AI that offers agents with tool access and email features: agents that touch your files and inbox are a potential attack surface. There's also a practical limit. Local models that run on ordinary home PCs usually can't match the quality of the biggest cloud models. If you want top performance, you'll end up back in the cloud via the OpenAI or OpenRouter connection. So how good Odysseus really is depends heavily on the model you pick and your own hardware.
Related videos
Sources
- Google News (SE) feber.se
- finance.yahoo.com finance.yahoo.com
- mezha.ua mezha.ua
- tbsnews.net tbsnews.net
Ask the AI assistant
Questions about this article, installation or background? Our assistant knows this page and runs on our own machine, not in someone else's cloud.
The Assistant is an AI (Gemma 4) and can make mistakes – please check important details against the source. Questions are used only to generate an answer and are deleted after one hour. Please don't enter any personal data.
Comments
What do you think? Already tried it? Write it here, no sign-up needed.
Comments are public. Links aren't allowed, and we delete abusive posts. You can delete your own comment at any time in this browser. Learn more in the Privacy Policy.