● Overview
The Shield private deployment puts the full AI inference stack on the customer's own server, so data never leaves the site. Offline inference means every AI function runs without an internet connection. It supports mainstream Linux and Windows servers with one-click Docker deployment; typical rollout takes no more than two working days.
● Core Capabilities
On-premise deployment
The inference engine runs on the customer's own hardware — nothing is uploaded to the cloud, so data security is enforced physically.
Offline inference
All AI selection, inspection and analysis functions run without internet access, suiting strictly air-gapped production environments.
One-click Docker deployment
Containerised deployment for mainstream Linux/Windows servers — short rollout, low operating cost.
Security audit compliance
Full operation logs and access auditing, meeting Level 2 of China's Multi-Level Protection Scheme (MLPS) for regulatory inspection.