The XE Genesis Node is built around the AMD Ryzen 9 8945HX – 16 cores, 32 threads – with 32 GB DDR5, 1 TB NVMe, and an optional NVIDIA A2000 GPU. All housed in the COOJ Sparrow MQ5 aluminium chassis.
One powerful platform. Every Genesis Node ships with the same high-performance base. Add an NVIDIA A2000 for GPU-accelerated AI inference and media transcoding.
16 cores, 32 threads. High-performance mobile workstation CPU designed for sustained, always-on operation. Handles Gemma 4 E2B and E4B models natively on CPU.
High-bandwidth DDR5 for running AI models, Docker containers, and self-hosting workloads simultaneously. Sufficient for Gemma 4 up to 26B MoE.
Fast boot and primary storage. Case supports additional internal or external storage expansion for media libraries, backups, and large datasets.
AI-Accelerated model adds an NVIDIA A2000 for significantly faster inference on 26B and 31B models, GPU-accelerated media transcoding, and enhanced multimodal processing.
XClaw dynamically selects the optimal Gemma 4 model based on task complexity and available hardware. All four variants are multimodal (text + image + video; smaller models also support audio). Context windows of 128K (E2B/E4B) or 256K (26B/31B).
Ultra-efficient edge model. Runs comfortably on CPU. Ideal for simple queries, quick responses, and low-power always-on use.
Balanced lightweight model. Excellent everyday performance on CPU or light GPU acceleration.
Mixture-of-Experts. High intelligence with excellent efficiency. Strong reasoning and agentic capabilities. Benefits significantly from GPU acceleration.
Highest capability. Complex reasoning, long-context tasks, coding, and advanced agent workflows. GPU strongly recommended for smooth 24/7 performance.
Designed for 24/7 uninterrupted operation. Quiet, efficient, and reliable. AKASA slimline PWM fan with aluminium unibody for thermal dissipation.
One-click deployment of curated Docker containers via XEOS. Bitcoin Core, Jellyfin, Nextcloud, Home Assistant, VPN servers, ad-blockers, and dozens more.
Local agentic AI powered by Gemma 4. Natural language control, intelligent system management, and a capability marketplace for extending your node's intelligence.
Family safety mode. Smaller, filtered Gemma 4 models for age-appropriate responses. Content filtering, parental controls, usage limits, and activity logging.
Custom operating system with file management, Docker apps, firewall control, blockchain wallet, and developer tools. Out-of-box experience, pre-configured.
Aluminium ITX chassis, 117 × 213 × 235 mm (~5.6 L). Ultra-compact. Beautiful on a desk, silent in a rack. Precision-machined, built to last.
The NVIDIA A2000 GPU in the AI-Accelerated model unlocks significantly faster inference and additional capabilities not practical on CPU alone.
