Build in public

The blog

Notes from running production AI agent systems. Architecture decisions, cost models, and the boring infrastructure most posts skip.

Latest··6 min read

How I run 23 AI agents on a Mac Mini M4 for a real agency client

The architecture behind the Strivve Media build — Mastra orchestration, local Llama 4 on a $599 box, and what production AI looks like when nobody's pretending it's enterprise.

Read post