Kuartal Mini PC · Node · Forge · Grid
Your own computer for AI.
Concept · Engineering targetsThe Kuartal personal-compute family is being designed to run KuartalOS, connect to Kuartal Cloud and support local workloads where hardware permits.
Pre-order coming soon — not available to purchase
Small footprint. Room to grow.
Compact dark housing with rounded corners, carbon-inspired texture, subtle green Kuartal lighting and front/rear connectivity. Planned expansion options include high-bandwidth external GPU support, subject to platform design.
With KuartalOS, the goal is a familiar desktop for everyday apps and an AI layer for local or cloud tasks.
Built to evolve.
Target features for engineering exploration, not confirmed shipping specifications.
| Operating system | KuartalOS (Linux-based roadmap); browser preview available separately |
|---|---|
| Compute platform | CPU, accelerator and board selection: to be announced |
| AI execution | Planned local model support plus optional Kuartal AI cloud routing |
| Memory and storage | Capacities and upgradeability: to be announced |
| Expansion | eGPU / high-bandwidth I/O targeted; interface and bandwidth not finalized |
| Networking and ports | Multiple USB, display and network connections envisioned; final count TBD |
| Chassis | Compact black/graphite housing, carbon-inspired finish and illuminated Kuartal emblem |
| Price, availability and warranty | To be announced; no orders or deposits accepted |
The lineup we're exploring.
Engineering targets / five device families. These configurations are earlier design targets and candidate components for prototyping—not a final bill of materials or shipping specification. Form factors, performance, ports, prices and availability remain unverified.
| Family | Candidate compute | Target memory / storage | AI / expansion direction |
|---|---|---|---|
| Node Lite | Intel N150 / N250 class | 16–32 GB · 512 GB–1 TB NVMe | Always-on gateway; integrated graphics; 2.5GbE target |
| Node | Ryzen 7 8845HS / Ryzen AI 7 350 class | 32–96 GB · 1–2 TB NVMe | Everyday workstation; USB4 / OCuLink eGPU candidates |
| Forge | Ryzen AI Max-class APU | 64–128 GB high-bandwidth memory · 2 TB NVMe | Local AI workstation; integrated accelerator, optional external expansion |
| Forge Max | Ryzen AI Max+ 395 candidate | Up to 128 GB unified memory · 2–4 TB NVMe | High-memory workstation for larger local models |
| Grid | Threadripper PRO / WRX90-class platform | 128–256 GB+ · expandable NVMe / NAS | Private tower or rack AI server; multiple GPUs / high-speed networking |
AI model size and speed will require repeatable benchmarks on the chosen build. Electrical, thermal, power, emissions and safety requirements must be validated before sale.
Hardware follows the software.
Try KuartalOS in the browser while the devices are in development.