# Chat # Implementations ## Deepseek Stack using Ollama & OpenWebUI ### Ollama (Manually) Start a server by first ``` ollama serve ``` (Default port is 11434). And then run your model (do just this to test a model.) ``` ollama run deepseek-coder ``` Stop the server by stopping the ollama service: ``` systemctl stop ollama ``` ### Or just run everything with Docker ``` docker compose down && docker system prune -af && docker compose build && docker compose up -d && docker logs -f openwebui ``` #### Notes on first time setup 1. Create admin account 2. Wait a moment for the UI to load, sometimes takes a long time (like legitimately 10 minutes) with my card.