Hardware

Private compute for Nythral-managed infrastructure.

Marketplace devices available for private AI runtimes, self-hosted infrastructure, edge automation, and client-owned managed deployments.

Live marketplace

4 offers

Order flow

Nythral ID

Provisioning

Managed setup

Hardware comparison

Choose the device by workload, not by specs alone.

Read the hardware guide
Device pathBest forMemory / workloadLimitations
Mac mini / compact nodeAlways-on private assistants, small local models, STT jobs, and lightweight automation.24-64 GB unified memory; best for compact models and low-noise office deployment.Not the right default for heavy multi-user inference or large model experiments.
Mac Studio / workstationFounder workstations, stronger coding assistants, creative AI, and larger local workflows.64-128 GB unified memory; better headroom for mid-size models and local experiments.Still not a replacement for dedicated multi-GPU serving when latency and concurrency matter.
Umbrel / self-hosting nodePrivate apps, self-hosted services, owner-controlled dashboards, and small automation.Sized for services and self-hosting first; model use depends on the specific unit.Best treated as an infrastructure appliance, not a raw AI inference box.

Private runtime fit

Devices are selected for private AI, automation, self-hosting, and infrastructure workloads rather than generic ecommerce inventory.

Provisioning ready

Hardware requests connect to Nythral Office so setup, access policy, telemetry, and client status can be tracked.

Account-bound purchase

Buy actions start in Nythral ID so the request attaches to the right customer workspace.