n8n-io/self-hosted-ai-starter-kit ? reverse-engineered prompt

Build me a simple local AI starter project that I can run with Docker Compose on my own machine.

I want it to bring up a self hosted n8n setup, plus local AI pieces like Ollama, Qdrant, and Postgres, so I can experiment with private AI workflows without sending data to outside services. Make it easy to start, with a clear .env file to copy and update, and let me choose the right setup for CPU, Nvidia GPU, AMD GPU, or running Ollama on my Mac separately.

When it’s running, I should be able to open n8n in the browser, see a ready to use example workflow, and chat with it right away. Please keep everything set up so a beginner can get started fast, and add any helpful notes in the README for first run, common issues, and how to change the Ollama connection if needed. If you need current docs or setup details, look them up online.