A Tiny Crew of Agents Running My Homelab
One main agent, two specialists, Gemma 4 E4B doing the thinking. Pi holding it all together.
Apr 17, 202612 min read91

Search for a command to run...

Series
What happens when LLMs stop being chat windows and start running real systems. Local models, MCP tools, multi-agent crews, and an audit trail for everything they touch.
One main agent, two specialists, Gemma 4 E4B doing the thinking. Pi holding it all together.

Exploring Local LLM Projects, Trade-Offs, and Insights

Most of us build homelabs to tinker, automate, and take control of our infrastructure. But somewhere between Docker containers, backups, and uptime monitoring, it becomes a lot to keep track of. I did
