Your Mac Already Has a Free AI Model Built in: Here’s How to Unlock It with Apfel
TL;DR
Apple exposes its local Apple Intelligence model to developers through the Foundation Models framework. On compatible Apple Silicon Macs it runs locally, without API keys, cloud calls, or separate model downloads. The open-source tool apfel adds a usable layer on top: a CLI, interactive chat, and an OpenAI-compatible localhost server, so existing scripts and SDKs can use the Mac as a local AI backend.
Nauti's Take
This is not a ChatGPT killer, and selling it that way would be silly. A small local model with a 4,096-token window is more like the screwdriver in the drawer than the workstation in the data center.
That is exactly why it matters: for small, repeatable tasks, you do not always want a cloud agent, an API key, and a privacy discussion. If apfel stays stable, local AI on the Mac finally becomes pipe-friendly.
Briefingshow
For many workflows, the winning model is not the most powerful one, but the one that is already available, local, and free to run. apfel turns Apple’s model into a terminal tool and a local API endpoint. That makes the Mac a small AI infrastructure layer for sensitive notes, scripts, and developer helper tasks.
Video
Sources