Use Cases
Curve Reality supports AI product teams and hardware builders creating systems where compute must be persistent, wearable, portable, or decoupled from the display layer.
AI Product Teams
Persistent AI systems that need continuity across sessions, devices, and models.
Build AI products that retain context over time instead of starting from zero every session. Combine local and cloud inference with a memory layer that preserves continuity, recall, and personalization as your architecture evolves.
Typical Requirements
- Persistent memory across sessions
- Compatibility with local and cloud models
- Low-latency edge inference with cloud-connected workflows
- A deployable hardware path for production pilots
How the Platform Helps
- Model-agnostic memory layer through iMember integration
- Jetson-class local inference with cloud-connected workflows
- Carrier boards for prototype-to-production hardware deployment
- Modular architecture that survives model and device changes
Wearable XR Compute
Body-worn compute for AR/VR systems that outgrow integrated headsets.
Build AR/VR systems where the headset stays light and the compute lives on the body. Separate optics from processing so you can iterate on displays without redesigning your compute stack.
Typical Requirements
- Low-latency video output to headset
- Continuous battery operation (2-8 hours)
- Compact, body-worn form factor
- Android (LineageOS) or Linux (L4T) support for XR runtimes
How the Platform Helps
- DisplayPort over USB-C for direct headset connection
- Hot-swappable M12 batteries for continuous runtime
- Compact carrier boards designed for wearable integration
- CB301 supports Android; CB302 supports Linux + JetPack
Mobile Robotics & Field Compute
Battery-powered edge AI for environments where wall power and racks don't exist.
Deploy AI-accelerated compute in mobile robots, drones, or field equipment where power efficiency and ruggedness matter more than rack space.
Typical Requirements
- NVIDIA Jetson-class AI acceleration
- Battery-powered operation
- Ruggedized or protected enclosure
- Camera and sensor interfaces
How the Platform Helps
- Orin Nano/NX modules with up to 100 TOPS
- Milwaukee M12 ecosystem for field-serviceable power
- Modular housing options for protection
- Dual MIPI CSI camera interfaces
- Persistent memory across missions — robots learn from every deployment instead of starting from zero
- Fleet-wide knowledge sharing through cloud memory sync
- Edge inference for real-time decisions, cloud integration for deep analysis
Industrial Training & Teleoperation
Wearable compute for operators who need both hands free.
Equip workers with wearable compute for training, remote assistance, or teleoperation. Systems that move with the operator, remain hands-free, and survive industrial environments.
Typical Requirements
- Hands-free operation
- Real-time video streaming
- Durable, field-serviceable design
- Easy battery swap during shifts
How the Platform Helps
- Wearable form factor keeps hands free
- Low-latency DisplayPort for AR glasses or HMDs
- Hot-swap batteries for continuous shift operation
- Industrial-grade M12 battery ecosystem
Research & Rapid Prototyping
Skip the custom hardware phase and go straight to software iteration.
Prototype next-generation wearable and spatial computing systems without building custom hardware from scratch. Iterate on software while the platform handles compute and power.
Typical Requirements
- Flexible I/O for sensor integration
- Quick iteration cycles
- Access to GPIO, I2C, SPI, UART
- Standard development toolchains
How the Platform Helps
- 40-pin GPIO header (Jetson devkit compatible)
- Production-grade boards reduce time-to-prototype and eliminate redesign before deployment
- Full JetPack SDK and L4T support
- NVMe storage for large datasets and models
Choosing a Board
CB301
Optimized for Android-based XR workflows on Jetson TX2 NX. Best for teams building wearable XR systems that require native Android and direct display output.
CB302
Designed for Orin-class performance, battery-powered AI, and deployment-focused systems. Best for robotics, field compute, and persistent AI applications running local and cloud-connected workflows.