One ISO. Boots from USB. Scans your hardware.
Downloads the right AI models automatically. Ready in minutes.
One universal image — works on any compatible machine. Around 2 GB. Contains the SteraOS base system and installer.
Use Balena Etcher (recommended), dd, or any USB imaging tool. Takes about 2 minutes.
Boot your machine from the USB drive. The installer runs automatically — it scans your hardware, selects the right models for your GPU, downloads them from the Stera platform, and installs everything.
Remove the USB drive. Your machine boots into SteraOS with AI models pre-loaded and ready to work. Sign in to connect to the Stera platform.
The installer detects your GPU and downloads the right model pack. More VRAM means larger, more capable models.
All model tiers include SteraOS, the inference engine, Talent Pod runtime, and the Stera Grid client. Models can be added or removed later from the Marketplace.
USB Boot
┌─────────────────────────────────────────────────────┐
│ │
│ 1. Hardware scan │
│ ├── GPU: model, VRAM, driver compatibility │
│ ├── CPU: architecture, cores │
│ ├── RAM: total capacity │
│ └── Disk: target drive, free space │
│ │
│ 2. Partition & base install │
│ ├── EFI boot partition │
│ ├── SteraOS root (Arch-based) │
│ └── Data partition (models, user data) │
│ │
│ 3. Model download (from platform.stera.se) │
│ ├── Detect GPU tier (Lite/Standard/Pro/Ultra) │
│ ├── Download model pack for tier │
│ ├── Verify checksums │
│ └── Install to /opt/stera/models/ │
│ │
│ 4. Configure & clean up │
│ ├── NVIDIA driver setup │
│ ├── SteraOS first-boot configuration │
│ ├── Remove downloaded installation files │
│ └── Set boot order │
│ │
│ 5. Reboot → SteraOS │
│ │
└─────────────────────────────────────────────────────┘
SteraOS installs as the primary operating system. For best performance, use a dedicated machine. Dual-boot with an existing OS is possible if you select a separate drive during installation.
The installer will ask you to confirm the target drive before writing anything. It will not touch other drives. All data on the selected drive will be erased.
The base system installs in about 5 minutes. Model download depends on your internet speed — 15 minutes on a fast connection, up to an hour on slower connections. The total is typically 20–40 minutes.
Yes. SteraOS includes a model manager. You can download additional models from the Marketplace at any time. If you upgrade your GPU, SteraOS will detect the new hardware and suggest larger models.
SteraOS is built on Arch Linux. NVIDIA GPU is required for AI inference. AMD GPU support is planned for a future release.
SteraOS is free to download and use. Optional licenses unlock Grid sharing, fine-tuning, and team features. One-time purchase, no subscription.
Community edition — full SteraOS, local models, no Grid sharing
Personal use — Grid sharing enabled, earn Joules from idle GPU
Professional — priority model updates, fine-tuning tools, extended support
Team — multi-user access, RBAC, centralized model management
All tiers include the full SteraOS operating system, local AI inference, and the Talent Pod runtime. No monthly fees. No per-token costs. You own the software and the hardware.
Free for the first 1,000 users. One universal ISO. Install on any compatible machine.
sudo dd if=steraos-latest.iso of=/dev/sdX bs=4M status=progressDownload Balena Etcher. Select the ISO. Select your USB drive. Click Flash.
Restart your machine and press F12, F2, or Del (depends on your motherboard) to enter boot menu. Select the USB drive. The Stera installer will start automatically.
The companion app for your Mac, Windows, or Linux computer. Control your Stera machine remotely — manage models, monitor GPU, browse your knowledge library, and run Stera Loom.
Manage your Stera machine from any computer on your network. Start/stop models, toggle GPU sharing, check status — all from your laptop.
Browse, search, and manage your AI's knowledge base. See what Stera has learned, remove entries, and track how knowledge grows over time.
Describe what you want, Loom builds complete applications for you. No coding needed. Uses your Stera Coder Pod — zero token costs. Learn about Loom →
Real-time GPU utilization, temperature, VRAM usage, and Grid earnings. See exactly what your Stera machine is doing at any moment.
Download the .dmg, drag Stera Dashboard to Applications. Requires macOS 13 Ventura or later.
Download the .exe installer. Works on Windows 10/11. No admin required — installs to user directory.
Available as .AppImage (universal) or .deb (Ubuntu/Debian). Run directly — no installation needed for AppImage.
Download and install on your Mac, Windows, or Linux computer.
Use the same Stera account you used on your Stera machine.
The Dashboard automatically finds your Stera machine on the local network. No IP address or port configuration needed.
Skip the installation. Pre-order a Stera machine — assembled, tested, and ready to power on.