# SelbsAI Server — Product Sheet

Version: 2026-08-29

## Product

SelbsAI Server is a complete, preconfigured x86-64 mini-PC appliance for running curated open-weight AI models on premises. It includes the compute, memory, encrypted storage, runtime, models, and local browser workspace.

## Available configurations

- Sai Start Server: 8 GB DDR5, 1 TB NVMe, selected 1B–4B quantized models, from EUR 499 net.
- Sai Personal Server: 16 GB DDR5, 1 TB NVMe, selected 3B–8B quantized models, from EUR 699 net.
- Sai Core Server: 32 GB DDR5, 1 TB NVMe, selected 7B–14B quantized models, from EUR 899 net.
- Sai Pro Server: 64 GB DDR5, 2 TB NVMe, selected 14B–32B quantized models, from EUR 1,499 net.
- Tracked EU-wide delivery: EUR 6.50.
- Typical preparation time: 5–14 business days, depending on configuration and optional Pre-Load.

## Included

- Local inference runtime and curated model.
- Private browser workspace at `http://selbsai.local`.
- Power supply and Ethernet cable.
- Laminated setup and recovery card.
- Installed-stack and model-provenance handover record.
- Customer portal for order status and support.

## Client requirements

No customer compute hardware is required for model inference. Users connect from a browser-capable PC or laptop on the same local network. SelbsAI sizes Core or Pro based on user count, workload, model class, and document volume.

## Product boundary

Core inference can run locally without a model-cloud account. Checkout, customer authentication, optional updates, support, shipping, and optional Pre-Load use defined online services.

Contact: orders@selbsai.com