Guides

Plain guides to local AI, private AI, and training a model on your own work, with measured results.

What is local AI?

Local AI runs on your own smartphone, tablet, or computer, not in a data centre. See what it does well, where it falls short, and what you need.

How to train an AI on your writing, emails, and documents. See what training changes, how to prepare examples, and a real run with raw outputs.

What private AI means, what stays on your device, what can leave it, and how to check the privacy claim of any AI tool yourself.

Memory and connectors bring facts to an AI. Training changes how it behaves. See what each one does, with an example, and when to use each.

What you can run on a Mac with 8 GB to 16 GB of memory. See why memory is the main limit, and how fast a small model can be.

How local AI works on a smartphone, what your phone needs, what works well, and how to check that an app is really local.

Measured scores of small local models beside Claude Opus 4.5, with the wins, the losses, and the method.