I upgraded my server just before Christmas (and before the price hikes) so I’ve got decent hardware. I’m running Proxmox with a variety of VMs. For the Ai stuff (which lives on a separate VLAN), I have spun up a new VM for each of the agents and my main one, Eve, orchestrates everything and does ops. Then one agent is a builder, one is a researcher, and one is my social media guy (although I’ve not done anything with that one yet). I’m using Openclaw and setting it up the first time was a bit tricky, but after that Eve does it for me. For the LLM I’m running a local one through Ollama, but the main one I’m using is gpt 5.4. If you have a chat GPT account you can log into Openclaw through the codex oauth, so I’m not wracking up big API bills. So far I’ve built an app for my wife’s business. A creative assistant to help me develop my ideas for YouTube videos, particularly around titles and thumbnails. And there’s a simple app that is basically project management and research. If it sounds like I know what I’m doing, I don’t. I just had a go and put in the time.