A1 gives your agent, files and long-term memory a home on your device, with cloud AI for demanding tasks. Local lightweight models depend on memory, quantisation and software support. Cloud requests are sent to your chosen provider.
Wired $249 / Wi-Fi $299. One-time hardware purchase. Cloud AI/API usage, third-party services and model licences may cost extra. Local inference does not incur cloud API token charges.
| CPU | Rockchip RK3576 Octa-core |
|---|---|
| NPU | 6.0 TOPS NPU, supports local inference |
| RAM | 4GB LPDDR4X |
| Storage | 64GB eMMC 5.1 |
| Network | Dual 2.5G LAN Ports |
| Wireless | Available with Wi-Fi variant; check selected configuration |
| Video | 1x HDMI 2.1 (4K@60Hz) |
| USB | 2x USB 3.0, 1x Type-C |
| Power | Typical 5W-12W, Peak <20W |
| Cooling | Passive Silent Cooling |
| Dimensions | 75x75x40mm |
| OS | Ubuntu 24.04 LTS (Pre-installed OpenClaw) |
| Model | Positioning | Price | AI Compute | Memory | Storage | Power | Best For |
|---|---|---|---|---|---|---|---|
| 铠盒AIBOX-A1 | Home Entry AI Agent Computer | $249 | 4GB LPDDR4X | 64GB eMMC 5.1 | Typical 5W-12W, Peak <20W | Home AI Hub, Personal AI Assistant, Student Programming | |
| 铠盒AIBOX-B1 | Industrial Lightweight AI Terminal | $699 | 16GB LPDDR5X | 128GB eMMC | DC 12V-36V | Factory Automation, Smart City Sensors, Remote Maintenance | |
| 铠盒AIBOX-C1 | NVIDIA Entry · Local AI Box | $799 | 67 TOPS (INT8) | 8GB LPDDR5 | 128GB NVMe SSD | 67 TOPS (INT8) | Local AI Assistant, Smart Home Hub, AI Algorithm Dev |
| 铠盒AIBOX-D1 | NVIDIA Mainstream · Performance AI Box | $1,899 | 70/100 TOPS | 16GB LPDDR5 | 256GB NVMe SSD | 70/100 TOPS | SME AI Gateway, Advanced Security, Multi-modal Agent Research |
| 铠盒AIBOX-E1 | AMD Flagship · Desktop Versatile AI Host | $1,989 | 55 TOPS (NPU) / 86 TOPS (Total) | 32GB DDR5 5600MHz (Up to 128GB) | 1TB PCIe 4.0 SSD (3x M.2) | 55 TOPS (NPU) / 86 TOPS (Total) | Content Creators, Indie Devs, Executive Office |
| 铠盒AIBOX-F1 | AMD Peak · Mobile Workstation AI Box | $3,299 | Total 126 TOPS, runs 70B-235B LLMs | 128GB LPDDR5X-8000 | 2TB PCIe 4.0 SSD | Total 126 TOPS, runs 70B-235B LLMs | AI Labs, LLM Developers, Enterprise Data Hubs |
| 铠盒AIBOX-G1 | NVIDIA GB10 Local AI Workstation | $8,999 | 1 PetaFLOP (FP4) | 4TB Enterprise SSD | 1 PetaFLOP (FP4) | AI development and experiments, Enterprise knowledge retrieval, Local model inference |