No products in the cart.

NeoEyes NE503
Edge AI Camera Platform
Run vision AI where video starts. NE503 combines high-resolution video input, local multi-model inference, containerized applications, and event-driven workflows for system integrators, algorithm developers, and OEMs.
Run vision AI where video starts. Turn 4K video input into on-device AI inference, application logic, and structured events.
Traditional IPC vs NE503
Traditional IPC Records. NE503 Turns Vision into Action.
Most IP cameras capture and stream video for external processing. NE503 runs AI models, application logic, and event workflows directly on the camera — turning visual input into structured events and local actions.
Highlights
Built Beyond Traditional IP Cameras









How NE503 Works
From Visual Input to Structured AI Events
Capture 4K AI-Ready Visual Data
Capture high-resolution visual input with Sony IMX678, Gen2 AI-ISP, and motorized optical zoom, giving AI models stable data from fixed-site deployments.
Capture AI-ready 4K input with Sony IMX678, Gen2 AI-ISP, and motorized zoom.

Run AI on the Camera
Powered by Hailo-15H, NE503 delivers 20 TOPS INT8 AI performance for local multi-model inference, reducing server dependency and enabling faster on-site decisions.
Run multi-model inference locally with 20 TOPS Hailo-15H AI compute.

Run Applications on the Camera
Deploy containerized applications that package AI models, workflow logic, and system connectors directly on NE503 for project-specific vision tasks.
Deploy custom models, workflows, and connectors as containerized apps.

Apply Logic Before Data Leaves
Filter model outputs, apply rules, and convert detections into structured AI events, so only meaningful results move forward to your systems.
Apply rules and convert detections into structured AI events.

Send Events to Your Existing Stack
Send structured events through MQTT, REST API, VMS integration, or local I/O to trigger alarms, update platforms, and connect with field devices.
Send events to MQTT, REST API, VMS, or local I/O for system actions.

Visual Input Quality
AI-Ready Visual Input for Real-World Scenes
NE503 combines a Sony IMX678 4K sensor, Gen2 AI-ISP, and motorized optical zoom to preserve the visual details AI models need in low light, high contrast, and variable working distances.
AI Compute
Run Multi-Model AI on the Camera
Powered by Hailo-15H 20 TOPS INT8 compute, NE503 runs multi-model vision workloads from a single camera stream — detection, OCR, pose, ReID, behavior analysis, and visual search — then turns AI results into structured events for your systems.

APP CONTAINER CAMERA
Run Complete Vision Applications at the Camera
Package AI models, business logic, user interfaces, and system connectors as one OCI application. Deploy and manage it directly on NE503 without rebuilding the camera firmware.

Package Your Workflow
Turn models, logic, and connectors into a deployable vision application.
View App Deployment Workflow →Deploy as OCI Images
Install apps from a container registry or upload a local application archive.
View App Deployment Workflow →Run in Isolation
Keep applications separated from core system resources with a sandboxed runtime.
View App Deployment Workflow →Upgrade Independently
Update models and applications without rebuilding the whole camera system.
View App Deployment Workflow →Supported Development Stack
Build with familiar languages, package as OCI images, and bring models from common AI frameworks.
C++ and Go support are planned and will be available in future releases.
OPEN Integration
Open Interfaces for Real-World AI Deployment
NE503 exposes video streams, structured AI events, APIs, and field interfaces so integrators can connect edge intelligence to VMS, IoT platforms, SCADA/BMS systems, backend applications, and on-site devices.

Connect Video Streams
Stream live video to VMS, NVR, or your own vision pipeline.
Control Device Services
Manage camera services, applications, and device behavior through APIs and SDKs.
Route AI Events
Send AI results and device alerts as structured outputs to your existing systems.
Link Field Devices
Connect AI events to alarms, access-control equipment, sensors, and industrial devices.
Dual-Board Architecture
Built for AI Workloads and Field Control
NE503 separates AI compute from field interface control. Vision applications can run, update, or restart on the camera while alarms, sensors, access control, and industrial interfaces stay connected through dedicated hardware layers.
AI apps can change. Field I/O stays connected.

FIELD DEPLOYMENT
Built for Fixed-Site Edge AI Deployment
NE503 is designed for real-world camera installations with PoE+ power and networking, IP67 protection, passive cooling, wide-temperature operation, and field-ready I/O.
Field Deployment
Ready for Real-World Fixed-Site Deployment
Install with one cable, operate outdoors across wide temperatures, run fanless around the clock, and connect AI directly to field devices.




Power and Network over One PoE+ Cable
Use one Ethernet cable to power NE503 and connect it to the network, simplifying fixed-site edge AI camera deployment.
Built for Harsh Outdoor Conditions
A sealed IP67 enclosure and -40°C to +60°C operating range support fixed-site deployment in exposed outdoor and industrial environments.
Fanless Cooling for 24/7 Edge AI
Passive cooling removes a fan-related failure point and helps reduce maintenance for fixed-site systems running continuously.
Connect AI Events to Field Devices
Connect alarms, sensors, access control, and industrial devices directly at the camera endpoint.




Full Technical Specifications
Complete hardware specs, interfaces, and deployment specifications.
Applications
Turn Visual Scenes into Local Decisions
NE503 runs AI inference, application logic, and event workflows directly at the camera endpoint — detecting, reading, verifying, and triggering actions while sending structured results to your existing site systems.
Documentation
Go Deeper Before You Evaluate
Platform Architecture
Understand the NE503 Software Stack
Explore how containers, AI runtime, Event Bus, SDKs, and hardware services work together.
Explore ArchitectureApplication Tutorial
Build a Person Detection App
Follow a real workflow from on-camera inference to structured events and local hardware actions.
Start the TutorialWHAT'S DIFFERENT
Traditional IP Camera vs Edge AI IPC
Compare both architectures and learn when local inference, structured events, and the NE503 are right for your project.
Read the BlogGet Started
Request an Evaluation Kit
Tell us about your project, and our team will follow up with evaluation availability, technical resources, and next-step guidance. Or contact us directly at sales@camthink.ai.
- Evaluation availability
- Technical resources
- Integration discussion
- Project-fit guidance
Submitted Successfully
Thank you! Your evaluation request has been submitted. We will get back to you soon.






