Open WebUI on Unraid - Your Own Private AI Assistant (Fully Local)

What if you could have your own private, Claude-style AI assistant running entirely on your Unraid server? No subscription, nothing ever leaving your house, and no one snooping on what you ask it. That's exactly what this month's app of the month: Open WebUI In this video I walk you through the whole thing from scratch. We install Ollama (making sure you grab the right version for your graphics card) put Open WebUI in front of it, and pick a model that actually fits your hardware. Then we get into the genuinely useful stuff: chatting with your own documents, so you can drop in all your household manuals and just ask questions instead of hunting through a 20 page PDF, and even having a vision model look at photos for you. Think of Open WebUI as your front door to Ollama, because once it's running, a whole world of other tools can plug into it too. This is a complete setup & use guide, so by the end you'll have your own private AI assistant up and running and actually understand how to get the most out of it. Key Links 🔗 🔗 Ollama: https://ollama.com 🔗 Open WebUI: https://openwebui.com 🔗 Ollama model library (check the latest models): https://ollama.com/library ▶️ Related videos on the channel 🔗 Local AI on Unraid (the intro, watch this first):    • Local AI on Unraid - The Stuff Nobody Tell...   🔗 A-Eye (local photo recognition with AI):    • This AI Analyses Your Photos, Renames Them...   🔗 Paperless AI (auto-tagging your documents):    • Install Ollama & AI Models for Paperless N...   ⏱️ Timestamps 0:00 - Opening - your own private AI chat 0:57 - What you'll be able to do by the end 1:14 - Watch my local AI intro first 1:35 - Open WebUI vs Ollama - interface vs engine 2:04 - Why isn't Ollama the app of the month? 3:02 - What you actually need (and the no-GPU option) 3:54 - Installing Ollama: pick the right version for your card 4:34 - Nvidia, AMD and the separate Intel container 5:44 - The Ollama template and where models are stored 6:16 - Keep-alive, idle power draw and the 5090 myth 8:21 - The --gpus=all parameter (and the AMD alt line) 8:51 - Installing Open WebUI 9:49 - Connecting Open WebUI to Ollama (and remote/Tailscale) 11:23 - Creating your local admin account 11:41 - Which model should I actually run? 12:08 - Params and quantization, kept simple 12:52 - The model cheat sheet by VRAM 13:32 - Leave headroom for your chat to grow 13:53 - Going further with which models to use 14:17 - Installing a model in Open WebUI 14:49 - Switching models and setting a default 15:24 - Chatting with your own documents (RAG) 16:07 - Building a Knowledge base and uploading manuals 16:46 - Using it with # in any chat 17:09 - Asking about a washing machine error ! 17:17 - Cross-checking the TV manual (and the citation) 17:45 - Add your own notes too 17:57 - Vision model - asking an AI about a photo 18:34 - Roundup and outro What will you build with Unraid? Get Started with Unraid in 15 minutes or less: https://unraid.net/getting-started