Stack recipes: tools that run together
The real question is rarely “how do I run Open WebUI” but “how do I run it with a model server and a vector database”. Each recipe below is one Docker Compose file with every image pinned to a release. We started each one on our own server and ran the checks listed on its page before publishing it, and a weekly job flags a pin that has fallen two major versions behind.
Ollama + Open WebUI
How do I run a private ChatGPT-style chat on my own machine with Ollama and Open WebUI?
Ollama, Open WebUI · started and checked 19 Nov 2026
Ollama + Open WebUI + Qdrant
How do I run local RAG, chat with my own documents, with Ollama, Open WebUI and Qdrant?
Ollama, Open WebUI, Qdrant · started and checked 19 Nov 2026
LiteLLM + Open WebUI
How do I put LiteLLM in front of my models and use it from Open WebUI?
LiteLLM, Open WebUI · started and checked 19 Nov 2026
n8n + Ollama + Qdrant
How do I run n8n AI workflows with a local model and a vector store, all self-hosted?
AnythingLLM + Ollama + Qdrant
How do I run AnythingLLM with a local Ollama model and Qdrant as its vector database?
AnythingLLM, Ollama, Qdrant · started and checked 19 Nov 2026