-
Hardware
Hardware Specs designed for 24/7 always-on Agent
Multi-scenario Intelligent Hub with passive sensing and proactive service, supporting daily scenarios for life, work, business, and industrial environments.
-
Model Support
LLM Support and Multimodal Sensing
Natively support up to 100B-level model on the edge Natively support multi-channel devices including cameras, microphones, on-device vision and voice analysis with high performance and low latency
-
Privacy
Privacy & Security
User data remains local, ensuring full privacy compliance with hardware-level confidential computing
A State-of-the-Art Edge Compute Platform
-
High-performance AI Compute
-
Massive Memory Bandwidth
-
Advanced Local LLM
Inference Support
The compute foundation for AI that gets things done
-
High-Performance Edge AI Compute
Optimized local inference engine with robust multi-modal model capabilities, minimizing latency and eliminating cloud dependency
-
Evolving Intelligence with long-term memory
AI workstations that help creators and teams organize context, retrieve memory, and execute workflows.
-
Scalable Orchestration Layer
Real-time edge intelligence for retail, logistics, robotics, factories, and operational systems.
Synchronizing Great Minds to Shape Agentic Infrastructure
A place to share hard-won ideas, turn frontier systems into reality, and build with an open community.