Post
Building my AI-assisted App Factory
How I am using a local Mac mini, open-source models, and automated pipelines to experiment with app ideas.
Written in January 2025. I have since stopped using OpenClaw, so read this as a snapshot of my setup at the time.
I have spent the last few months turning my living room into a quiet laboratory. At the centre of it sits a Mac mini running Ollama, Open WebUI, and a growing set of local services I collectively call the App Factory.
Why local?
Cloud GPUs are brilliant, but I prefer the latency and privacy of my own metal. The Mac mini stays on, sips power, and never asks me for a credit card surprise.
The stack so far
- Ollama for local LLMs
- Open WebUI as a chat interface and prompt playground
- OpenClaw / Jarvis as a control bridge
- Astro for the site you are reading now
What comes next
I am wiring the blog drafting system into this pipeline. When it works, an idea can travel from voice memo to approved post without touching a cloud service.
This is a build log, not a launch announcement. I will share what breaks and what actually ships.