PewDiePie's Odysseus AI workspace outperforms Open WebUI and AnythingLLM in self-hosting test

I tried Open WebUI, AnythingLLM, and Odysseus to self-host my AI workflow, and only one delivered

Self-hosted AI has moved past the chatbot phase. There's a whole category of tools now that wrap around your model of choice and give it a proper home, with memory, docs, tools, agents, and whatever else. All you have to do is plug it into your runner and the workspace handles the rest. In theory one of these is all you need to cover most of what you'd do with cloud AI, just without the subscription and without your chats on a random server. I've been rotating through a couple of these AI workspace tools and they're all doing the same thing in very different ways... Want to stay in the loop with the latest in AI? The XDA AI Insider newsletter drops weekly with deep dives, tool recommendations, and hands-on coverage you won't find anywhere else on the site. Subscribe by modifying your newsletter preferences! AnythingLLM turned my local notes into something with memory It's built for the parts your model runner leaves out AnythingLLM is made by Mintplex Labs and is fully open source. The way it works is that it sits on top of your existing model runner instead of replacing it, so LM Studio, Ollama, llama.cpp, or vLLM keeps doing the actual inference, and AnythingLLM wraps a full workspace around it. You can also point it at cloud APIs like Anthropic or OpenAI if you want to mix in a frontier model on the same interface. The main reason I picked it up was persistent memory, and it delivers. A background process reads through your recent chats every so often, pulls out facts worth keeping, and re-injects them into future conversations so your local model actually knows what you told it a few weeks ago. There's a workspace scope and a global one, and honestly this is a much nicer setup than trying to wrangle memory through MCPs or custom prompts inside LM Studio. While I haven't taken full advantage of AnythingLLM yet since it's primarily a tool I use for memory, Workspaces are worth a mention. Each one is its own isolated bubble with its own docs, memory, model, and settings, so you can have a research workspace running a local model and a coding workspace pointed at Claude API and nothing crosses over. Then there's the agent side, which lets a model actually go do things like browse the web, run SQL, save files, or use MCP servers, and it activates through @agent inside any chat. Overall, it's a very useful tool, albeit a little stripped down for a workspace. AnythingLLM See at AnythingLLM Expand Collapse Odysseus is doing more than any workspace should The one built by PewDiePie Odysseus is the newest of the three by a wide margin and it's also the strangest. It's a self-hosted AI workspace built by PewDiePie and it launched a month ago at the time of writing. In roughly a month it's picked up over 77k GitHub stars, which for a project that came out of nowhere is a lot. It's Python and FastAPI with ChromaDB for its vector store, and it hooks into Ollama, llama.cpp, LM Studio, vLLM, or cloud APIs if that's your thing. Setup takes a minute if you go the Docker route, which is what I did, but the payoff is worth it. The interface itself is really pleasant to look at and I don't know if I've ever said that about a self-hosted anything before. My favorite thing in here is the Cookbook, which scans your machine and recommends models that'll actually run on your hardware, pulling from a catalog of over 270 options. So instead of guessing whether a model will fit into your VRAM or downloading a 12B just to find out it thrashes your system, the Cookbook tells you upfront and lets you serve the pick straight from the same screen. Chat is honestly the least interesting thing in there. Still, every conversation has a Notes panel attached to it, which gives me NotebookLM vibes in the best way, so you can jot down where you left off, save the useful bits from a response, and keep your scratch work next to the actual model without leaving the tab. There's also a to-do manager that the AI can actually see and interact with, so it'll add tasks, prioritize them, or work through them with you conversationally, which is a genuinely useful thing to have in the same window as your chat. And another really weird tool to find here, but I've been playing around with it nonetheless, is the image editor...in an AI workspace of all things. This is PewDiePie's stated attempt at his own Photoshop. Odysseus See at Odysseus Expand Collapse Open WebUI is the reliable one Steady and capable, but boring Open WebUI is the elder statesman of the three, sitting past 350 million downloads and comfortably one of the most-installed self-hosted AI interfaces out there. It's positioned as a full ChatGPT-style workspace you run yourself, and like AnythingLLM, it's backend-agnostic, so it plugs into Ollama, LM Studio, vLLM, LocalAI, or any OpenAI-compatible API. Docker is the main install path with a single command. The chat itself is clean but boring; I mostly like that each conversation has a Notes workspace attached to it with a rich editor where you can draft freely, let the AI rewrite selected chunks, and then feed the whole note back into a chat as context. There's also Channels, which is where things get more interesting than a normal chat interface tends to go. Channels are shared spaces, kind of Slack-style, where you and your teammates and multiple AI models all live in one timeline. You tag a model with @ to pull it into the conversation and it replies right there in the thread, so you can have GPT draft something, tag Claude to critique it, and pin it. Persistent memory is in here too and works automatically across chats, so your context will carry forward. Plugins are its other real strength, with Filters, Actions, Pipes, Tools, and Skills, plus MCP and OpenAPI tool server support, so you can extend it pretty far if you want to. Open WebUI See at Github Expand Collapse My pick isn't the one I expected AnythingLLM does its job well but I do think it feels a tad stripped down for something calling itself a workspace. Open WebUI is steady and covers everything you'd need, but it's a little boring, and there's nothing wrong with boring, it's just not what pulls me in in this case. Odysseus came out of nowhere for me, and I love the mix of tools it ships with, the interface, and how much of it feels like it was created with character. If all three demanded a subscription, Odysseus is the one I'd consider paying for.

I self-hosted PewDiePie's Odysseus AI workspace, and it's surprisingly brilliant

A genuinely powerful privacy-focused AI workspace made by a YouTuber? Get out of here. AI coding and workspace platforms are a dime a dozen these days, but I couldn't help but be surprised to see that PewDiePie has developed an AI workspace called Odysseus. The YouTuber is, perhaps surprisingly, pretty hot on powerhouse Linux and AI tools. If you're not familiar with an AI workspace, it's a platform that allows AI to interact with other tools and data; it could be a coding workspace or a way to have AI help out with your documents and emails. What initially caught my interest about Odysseus is its privacy-first approach. It's designed to be self-hosted and is easy enough to compile into a Docker image to run on my DIY NAS. This has the obvious benefit that emails, documents, personal information, and anything else you share with your large language models of choice remain yours, not saved and potentially exploited by a third-party company. As such, Odysseus is geared towards self-hosting your own AI models as well, ensuring that absolutely no data leaves your premises. One of the first creature comforts is the "Cookbook," which surfaces the wide range of models and their quantized variants hosted on Hugging Face, and ranks them by compatibility and size relative to your GPU. As someone who's dabbled in LM Studio and Ollama for hosting some small local models, this is a familiar and more user-friendly way to test and manage models on your own hardware. However, for someone who balks at the current price of cutting-edge RAM and GPUs, I'm not in a position to run some of the most powerful models on my own hardware. Thankfully, Odysseus supports bringing your own inexpensive cloud-hosted models via OpenRouter or similar API key setups for an acceptable halfway house. The data stays with me; I just have to trust that API endpoints aren't logging all my chats. Once you're set up with your model(s) of choice, the real fun can begin. As someone forever digging through technical specifications and getting my head around emerging tech topics, I've become a fan of the deep research tools you'll find at the likes of ChatGPT and Gemini. While I don't trust the results outright, deep research can be a great starting point for diving deeper into a project or for grabbing a summary of a well-established field. But as handy as these are, I'm not prepared to lay down $20 to increase the limits to some still-often-restrictive 25 or so queries a month. Odysseus supports deep research as part of its bring-your-own-model approach, searching and sifting through multiple web results, distilling the information, and outputting it as a nicely formatted web document for you to read. Having greater control over your models is surprisingly beneficial here; you can run the same research tasks with two or three models to make sure one of them didn't omit or misinterpret anything. But deep research is just one in a familiar toolset. Odysseus AI chats support web browsing; you can attach files and images to pass to supported models, and even inject prefix and suffix prompts to fine-tune your AI's responses. Bottom line, you're not missing out on tools from chat interfaces like ChatGPT or Gemini, aside from image generation, perhaps. In fact, Odysseus has some very interesting tools of its own. One of the workspace's more interesting tools is Persona, which essentially lets you save recallable system prompts for conversations. The defaults are quite goofy, like impersonating Socrates, but I've had plenty of success setting up personas specifically to look for spelling and grammar errors and help remove passive voice from my article drafts. Very handy, and this will feel right at home for those familiar with Custom GPT or Gemini Gems. You can also introduce several personas into a single group chat, allowing multiple AI models with their own system prompts to interact with one another's responses as the chat progresses. I haven't come up with a particularly handy use case for this, but it could be a good tool for cross-checking ideas or catching errors when solving complex problems. Odysseus also has scheduled tasks, automated email summaries, and a built-in calendar to help you stay on top of your life and work. A feature I have found particularly useful is document co-editing, which allows agents to work on markdown documents, code files, CSV spreadsheets, and even PDFs alongside you. This is quite helpful for drafting and reviewing your work or bug-checking short code snippets. You can go further with workspaces, which give your agent access to a dedicated folder and all its contents, rather than the more limited documents stored in its SQLite database. Though I'll admit I haven't tried hooking this up to a coding project yet, I imagine it could work as a Claude/Codex/OpenCode alternative for some tasks. Workspaces aren't quite as dangerous as, say, OpenClaw, as Odysseus is built to run in a Docker container with carefully scoped access to specific locations, bound by Docker and Linux user permissions. In other words, it can't read your entire file system; it can only read the folders you let the container access. However, shell commands can still reach out of the folder you pick for a workspace to hit other files in Odysseus' file path, so tread carefully. Odysseus also supports IMAP and SMTP email clients, CalDAV calendars, MCP Tool servers, and integration with other API services, allowing you to greatly extend your agents' reach across other tools. Sadly, Google Calendar isn't supported, and I couldn't get email summarization to actually do anything. There's also an image editor and a to-do list feature I haven't played around with yet. There's still a fair bit more of the system for me to learn before it feels like a true personal assistant. Speaking of learning, one of the things I found most fascinating about my time with Odyseus is its "self-learning" capabilities. Agents can automatically keep notes during their interactions with you, forming memories of your identity, preferences, and other important aspects of you and your work. The workspace quickly pinned me down as an editor and began making notes like "values scientific accuracy in technical writing" and "has strong technical knowledge of battery physics and power bank specifications." How sweet of it to notice. While it's fascinating to see what your assistant picks up on, you can also add, delete, or edit these memories to keep things on the right path. On those lines, agents will also create and refine their own skill set to help them solve tasks, or you can download or write skills for them. Self-developed skills have a confidence score indicating how well they seem to work at accomplishing a task, which seems like a very smart approach to self-learning and improving over time. Now, Gemini, ChatGPT, and others have an element of personalized learning too, but Odysseus feels smarter and far more transparent about what it's doing. Plus, you know for a fact that when memory data is deleted, it's gone for good. Furthermore, the development of self-reinforcement skills and memory refinement undoubtedly make the platform much more powerful and tailored to your specific needs the more you use it. I can't say I feel that way about my countless hours with ChatGPT. After spending several weeks with Odysseus, I've glimpsed the not-too-distant future of far more powerful and useful AI tools. While it would be nice to have deeper support for a wider range of office document types, easier cloud calendar integration, and workspace file management, the platform can already do a huge amount, and code contributions are only ramping up. On top of that, the self-learning, unique chat features, and bring-your-own-agent approach are a breath of fresh air in a sea of increasingly similar "agentic" platforms. The privacy angle is a really nice added bonus, too. Honestly, I'm impressed by what PewDiePie (and his AI agents) have built here, so much so that I intend to keep using Odysseus as my primary AI workspace. That's certainly not something I thought I'd say when I started this journey.