Sighthound Compute Hardware

Smarter Video Analytics Starts at the Edge

Run real-time AI on-device; no cloud, no delays. Sighthound Compute Hardware helps security teams, parking operators, and enforcement agencies turn video into instant decisions.

Want to test it? Ask us about trial kits.

About Our Hardware

Built for the Real World. Ready for Your Software.

Sighthound Compute devices combine powerful AI processing with rugged industrial design. Whether you're retrofitting an existing IP camera network or deploying new infrastructure, our edge devices provide fast, flexible, and secure computer vision, right where video is captured.

Better Tech

  • IP67–IP69K protection against dust, water, and weather

  • High-resolution imaging sensors

  • Passive thermal design; fanless operation (camera only)

  • PoE-capable devices for quick and easy installation, no separate power lines required

  • Built-in NVIDIA GPU with at least 256 CUDA cores (up to 1024 in Orin variants)

Better Software

  • Linux OS, Docker-ready, and OTA configurable

  • Run Sighthound ALPR+ or your own Linux-based apps

  • Compatible with existing RTSP/IP cameras

  • Supports popular AI development frameworks

  • API-ready for third-party system integration

Better Product

  • Designed and assembled in the USA

  • Short lead times, reliable inventory

  • Responsive technical support team

  • Proven in parking, public safety, retail, and logistics

Advanced Edge Computing & AI Devices

Sighthound Compute Camera 4

High-Resolution AI Camera with Built-In Compute

The Sighthound Compute Camera 4 combines a 4K Sony STARVIS imaging sensor with an onboard GPU, enabling real-time ALPR+ and object detection without external servers. Designed for harsh conditions and low-latency analytics, it’s ideal for fixed installations in parking, perimeter, and enforcement applications.

Key Features:

  • 8.3MP Sony STARVIS IMX 334 Sensor

  • Motorized zoom lens options: TL410P (4–10mm) or TL1250P (12–50mm)

  • Up to 4TB NVMe SSD storage

  • IP67-rated, passively cooled

  • Runs Sighthound ALPR+ or Docker-based custom apps

  • PoE+ powered, OTA service-ready

download datasheet

Sighthound Compute Node 5

Plug-and-Play AI for Parking, Security, and More

The Sighthound Compute Node 5 is a rugged, drop-in edge compute device that converts your existing IP cameras into intelligent enforcement tools. It delivers high-accuracy ALPR+ and vehicle analytics, no cloud required.

Key Features:

  • Real-time ALPR+ with Make, Model, Color, Generation (MMCG)

  • Works with any IP camera (RTSP)

  • Supports up to 4 streams in Super Mode (40W)

  • Built on NVIDIA Orin NX or Nano

  • Docker-based, API-ready for custom integrations

  • No infrastructure replacement required

Download Datasheet

Use Cases

  • Parking Management & Enforcement

    Parking Management & Enforcement

    Detect overstays, manage EV sessions, and match license plates to payment zones, all without any cloud infrastructure.

  • Gate Access & Automated Entry

    Gate Access & Automated Entry

    Trigger access control systems using license plate detection and vehicle verification.

  • Traffic & Transportation Analytics

    Traffic & Transportation Analytics

    Understand vehicle flow, enforce restricted lanes, and conduct safety studies at the curb.

  • Retail Security

    Retail Security

    Monitor storefront lots, pickup zones, and loading docks. Trigger alerts for after-hours activity or unauthorized access.

  • Campus Safety

    Campus Safety

    Secure student lots, staff entrances, and restricted zones. Detect permit mismatches or suspicious vehicle behavior in real time.

  • Public Safety & Law Enforcement

    Public Safety & Law Enforcement

    Process evidence, track suspects, or enforce time-based rules with rugged, portable edge units.

Built to be Flexible

Sighthound Compute Board

Run What You Want, Where You Need It

Both the Compute Camera and Compute Node run on our 3rd-generation Sighthound Compute Board. It combines a quad-core ARM CPU with a powerful NVIDIA GPU, giving you the freedom to run advanced AI models directly on-site, without any cloud dependency.

NVIDIA

Use it to run:

  • Sighthound ALPR+

  • Your own TensorFlow, PyTorch or custom AI models

  • Open-source object detection

  • Third-party enforcement or analytics tools via Docker

Specs Highlights:

  • Quad-core ARM CPU

  • Embedded NVIDIA GPU with up to 1024 CUDA cores

  • NVMe SSD storage (up to 4TB)

  • Linux OS

  • OTA service updates

  • USB, CAN, RS-485 I/O ports

Frequently Asked Questions

For more frequently asked questions about Sighthound Hardware and other Sighthound solutions, check out our Frequently Asked Questions page.

If you do not see the answer you are looking for, contact our experts.

Sighthound Hardware

  • Unlike server-based systems, Sighthound processes data locally, offering faster response, lower bandwidth usage, and full offline capabilities.

  • Yes. It supports standard RTSP streams and works with most IP cameras, no need to rip and replace.

  • No. Everything runs on-device, even ALPR+ and vehicle recognition. You can integrate with cloud or VMS systems later if desired.

  • Yes. Our US-based team provides responsive support, and we offer guidance for both installation and API integration.

  • Sighthound hardware utilizes the power of edge computing by processing data and running AI algorithms directly on the devices themselves, at the edge of your network. This significantly minimizes latency, improves responsiveness, and reduces reliance on constant network connectivity. With Sighthound Hardware, you can make critical decisions in real-time, ensure continuity of operations even in unreliable network conditions, and effectively manage and analyze your visual data without overwhelming your network infrastructure.

Become a partner

Sighthound empowers thousands of end users with computer vision solutions.

Learn about our partnership opportunities.