Build

Spin up your fit.

Drag, tap, and watch the machine take shape. No signup. Nothing uploaded — your selections stay on this device until you send them.

Accounts, not people using it at once.

Your machine · live

7 of 7 set
Accelerator memory planning baseline

~24–64 GB

Index memory ~8 GBStorage ~375 GB

Planning estimates, not a quote — your documents, models, and usage move these numbers.

Estimated fit: Solo

A focused private AI workstation for a small team.

What’s setting this

  • A compact footprint — everything here fits the smallest build.

Room to grow: nudge Demand or Model and watch when the fit steps up.

How this works, and what it is built on

Every answer can set a floor under the starting tier. The highest floor wins — and we show you which answers set it, and the one change that would lower it.

The tier boundaries are real: they come from what each build can physically hold, and from the bands published on Solutions. The memory figures are an engineering model, not a measurement — our published benchmarks cover throughput and latency, not memory occupancy per model class, so the envelope is our best estimate pending a measured run. We would rather say that than let a computed number imply more certainty than we have.