Local AI machines,
sized to your team.
Four Apple machines, from a Mac mini to a Mac Studio, each with the LocallyAI harness pre-installed — here is exactly what's inside, what it runs and what it costs installed.
16–128 GB unified · Serves 1–10+ · Fast and Smart lanes + Whisper, on the box · Unlimited seats
The Compact
Solo practitioners · 1–2 people
For one or two people who mainly want consults and meetings turned into finished notes and letters, drafted in your templates.
- Machine
- Apple Mac mini
- Memory
- 16 GB unified
- Serves
- 1–2 people
- Installed from
- $2,500 ex GST
16 GB, allocated
The Office
Small teams · 2–5 people
The most popular starting point. Both model lanes fit, so the router can send the hard questions to the Smart lane.
- Machine
- Apple Mac mini or AMD AI 395+
- Memory
- 32–48 GB unified
- Serves
- 2–5 people at once
- Installed from
- $5,000 ex GST
32 GB, allocated
The Practice
Whole offices · 5–10 people
For an office where several things happen at once — a consult transcribing in one room while a contract summary drafts in another. Reads full agreements and complete histories in a single pass.
- Machine
- Apple Mac Studio or AMD AI 395+
- Memory
- 64 GB unified
- Serves
- 5–10 people, several jobs at once
- Installed from
- $8,000 ex GST
64 GB, allocated
The Firm
Document-heavy practices · 10+ people
Our largest standard machine, for firms buried in discovery or clinics with many practitioners. Both lanes run at full quantisation with room for background jobs.
- Machine
- Apple Mac Studio or AMD AI 395+ (128 GB), top spec
- Memory
- 96–128 GB unified
- Serves
- 10+ people, heavy simultaneous use
- Installed from
- $10,000 ex GST
128 GB, allocated
The Portable
On the move, or serving the room — depends on what the day needs
The same harness on a laptop. On your desk it serves the office like any of the machines above; picked up, it's a private AI that travels — between rooms, out to site, down to court — with no network and no box to connect to.
- Machine
- Apple MacBook Pro
- Memory
- 36–64 GB unified
- Serves
- The whole office plugged in · one person on the move
- Installed from
- Quoted to the spec you pick
The Harness, on the laptop itself
We would rather sell you the right machine than the big one.
Every tier runs the same Fast and Smart lanes — bigger machines simply run them at higher precision and fit more models at once. We typically run models at Q6–Q8, occasionally Q4 where a system needs it.
Default model options (other models on request)
The one roster, for every tier on this page. Every machine also keeps memory in reserve so it stays fast on its busiest day — and if you prefer AMD workstation hardware, it's available on request.
A note on the models
The models named on this page are what we install today — better ones keep arriving. Our latest recommendations are easily accessed and downloaded straight through the browser, provided the machine has an internet connection.
Every install ends with a health check across every service on the box, and the same check runs again at every visit.
Book a 20-minute chat.
We'll tell you honestly whether this suits your business — and what it would cost.