1–20 users. Individual machines. Each one sovereign.
┌──────────────┐ ┌──────────────┐ ┌──────────────┐
│ Stera │ │ Stera │ │ Stera │
│ Machine #1 │ │ Machine #2 │ │ Machine #3 │
│ │ │ │ │ │
│ ┌──────────┐ │ │ ┌──────────┐ │ │ ┌──────────┐ │
│ │ RTX 4090 │ │ │ │ RTX 4090 │ │ │ │ RTX 4070 │ │
│ │ 24 GB │ │ │ │ 24 GB │ │ │ │ 12 GB │ │
│ └──────────┘ │ │ └──────────┘ │ │ └──────────┘ │
│ │ │ │ │ │
│ SteraOS │ │ SteraOS │ │ SteraOS │
│ Models │ │ Models │ │ Models │
│ Knowledge │ │ Knowledge │ │ Knowledge │
└──────┬───────┘ └──────┬───────┘ └──────┬───────┘
│ │ │
└─────────────────┼─────────────────┘
│
┌────────┴────────┐
│ GPU Pool │
│ Manager │
│ (free, opt.) │
│ │
│ Unified VRAM: │
│ 60 GB pool │
│ Basic dashboard│
└─────────────────┘
┌──────────┐ ┌──────────┐ ┌──────────┐
│ Your │ │ Your │ │ Your │
│ Laptop │ │ Desktop │ │ Phone │
│ │ │ │ │ │
│ SteraLink│ │ SteraLink│ │ SteraLink│
└──────────┘ └──────────┘ └──────────┘
Each Stera machine is a complete AI system. It has its own GPU for inference, its own storage for models and accumulated knowledge, and its own instance of SteraOS managing everything. You buy one. You plug it in. It works. There is no setup wizard, no cloud account, no subscription.
A single machine is enough for an individual — a developer, a writer, a researcher, a small business owner. It runs your models locally, learns from your work patterns, and operates autonomously on objectives you define. When you need more capability, you add another machine. Each one is independent. Each one is sovereign.
When you have three or more machines, you can optionally install the GPU Pool Manager — free, open-source software that unifies their GPUs into a single shared pool. A 24 GB machine plus a 24 GB machine plus a 12 GB machine becomes a 60 GB pool. Larger models that could not fit on any single machine can now run across the pool. A basic dashboard shows utilization, health, and model allocation.
Per Machine
─────────────────────────────────────────
Hardware GPU + CPU + RAM + SSD + PSU
Operating Sys. SteraOS (pre-installed)
Models Commander + specialist models
Knowledge Local persistent storage
Perception Vision + hearing
Autonomy Task decomposition + execution
Interface Voice, text, SteraLink app
With GPU Pool Manager (free, 3+ machines)
─────────────────────────────────────────
Unified VRAM Combined GPU pool
Dashboard Basic utilization + health
Load Balance Automatic inference routing
Model Sharing Run larger models across pool
Every machine ships ready to work. SteraOS is pre-installed. A commander model and essential specialist models are included. The machine perceives through camera and microphone, reasons about objectives, and acts autonomously. You interact through voice, through text, or through the SteraLink app on your existing devices.
Stera Machine
─────────────────────────────────────────────
16 GB GPU ······················· from €599
24 GB GPU ······················· from €899
48 GB GPU ······················· from €2,499
GPU Pool Manager ·················· Free
SteraLink App ·················· Free
Monthly fees ·················· €0
Per-token costs ·················· €0
One-time purchase. No subscription. No per-token costs. No seat licenses. The machine is yours. The intelligence it builds is yours. The cost of running it after purchase is electricity.
The sole & small tier is designed to grow naturally into the mid tier. When your team exceeds 20 people, when you need department-level budgets and access control, when you want software pods on a centralized server instead of individual hardware — the transition is seamless. Your existing machines can join the new GPU pool. Your accumulated knowledge carries forward. Nothing is lost.