SAFR System Requirements

Windows Requirements

Product Description Minimum Requirements Recommended Requirements
Desktop Client, Legacy Edition An adminstration console for SAFR. Use the Desktop Client to connect to a camera, process video, detect faces, and submit images to SAFR for recognition.
  • Windows Server 2016 or Windows 8.1, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i5-8259U or AMD Ryzen 7 2700X
  • NVIDIA GTX 1050 Ti or Quadro P1000
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • NVIDIA driver 418.96+ for GPU-enhanced performance1
  • 16GB RAM
  • 1.5GB available storage
  • Supports 2-3 4K cameras2
  • Supports 4+ 1080p cameras2
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen 7 2700X
  • NVIDIA GTX 2070 Ti or Quadro P4000
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • NVIDIA driver 418.96+ for GPU-enhanced performance1
  • 16GB RAM
  • 1.5GB available storage
  • Supports up to eight 4K cameras2
  • Supports 9+ 1080p cameras2
Desktop Client, CUDA 11 Edition A version of the Desktop Client that supports CUDA 11, allowing it to fully leverage the GPUs of NVIDIA cards with Ampere or newer architecture. (i.e. NVIDIA GEFORCE RTX 3*** or later)
  • Windows Server 2016 or Windows 8.1, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i5-8259U or AMD Ryzen 7 2700X
  • NVIDIA GEFORCE RTX 3060+
  • NVIDIA driver 456.38+
  • 16GB RAM
  • 1.5GB available storage
  • Supports 2-3 4K cameras2
  • Supports 4+ 1080p cameras2
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen 7 2700X
  • NVIDIA GEFORCE RTX 3060+
  • NVIDIA driver 456.38+
  • 16GB RAM
  • 1.5GB available storage
  • Supports up to eight 4K cameras2
  • Supports 9+ 1080p cameras2
Desktop Lite A version of the Desktop Client with fewer features. Desktop Lite has all the features of the Legacy Edition except for body detection, GPU acceleration, VMS extensions, and SAFR Actions.
  • Windows Server 2016 or Windows 8.1, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i5-7260U
  • 16GB RAM
  • 0.5GB available storage
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i7-8750H
  • 16GB RAM
  • 0.5GB available storage
SAFR Actions Actions allow you to create and manage responses to event triggers. Deploy them to unlock a door, turn on a light, send an alert, record data for reporting, or any security response to fit the use case.
  • Windows Server 2016 or Windows 8.1, or later versions of either
  • Intel Core i3-4340 or AMD Ryzen 7 1750
  • 1GB RAM
  • 1GB available storage
  • Windows Server 2016 or Windows 10, or later versions of either
  • Intel Core i5-726OU or AMD Ryzen 7 1950
  • 1GB RAM
  • 1GB available storage
SAFR Server3 The trusted engine of SAFR solutions. The SAFR Server includes: the facial recognition server, identity database, recognition event server, event archive, report server, and remote video feed administration servers.
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen TR 2700X
  • NVIDIA Quadro P2000
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • NVIDIA driver 418.96+ for GPU-enhanced performance1
  • 16GB RAM
  • 1TB available storage
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen TR 3700
  • Quadro RTX 5000 or Tesla T4
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • NVIDIA driver 418.96+ for GPU-enhanced performance1
  • 32GB RAM
  • 1TB available storage
SAFR Server3, CUDA 11 Edition A version of SAFR Server that supports CUDA 11, allowing it to fully leverage the GPUs of NVIDIA cards with Ampere or newer architecture. (i.e. NVIDIA GEFORCE RTX 3*** or later)
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen TR 2700X
  • NVIDIA GEFORCE RTX 3060+
  • NVIDIA driver 456.38+
  • 16GB RAM
  • 1TB available storage
  • Windows Server 2016 or Windows 10, or later versions of either
  • .NET Framework 4.6.2 or later
  • Intel Core i9-7980XE or AMD Ryzen TR 3700
  • NVIDIA GEFORCE RTX 3060+
  • NVIDIA driver 456.38+
  • 32GB RAM
  • 1TB available storage

1 = The Legacy Edition supports GPU acceleration for the following cards: Tesla T4, Quadro RTX 6000, Quadro RTX 5000, Quadro RTX 4000, Quadro P6000, Quadro P5000, Quadro P4000, Quadro P3000, Quadro P2000, Quadro P1000, GeForce RTX 2080 Ti, GeForce RTX 2080 Super, GeForce RTX 2080, GeForce RTX 2070, GeForce RTX 2060 Super, GeForce RTX 2060, GeForce GTX 1080 Ti, GeForce GTX 1070 Ti, GeForce GTX 1660 Ti, GeForce GTX 1650 Super, GeForce GTX 1060, and GeForce GTX 1050 Ti. If you have a newer card, install the CUDA 11 Edition instead.

2 = Number of cameras is based on an average of five visible faces in a 4K resolution camera view, running at 15 frames per second. Using fewer faces per camera and lower resolution will enable support for more cameras.

3 = Installed as part of the SAFR Platform installer.

macOS Requirements

Product Description Minimum Requirements Recommended Requirements
Desktop Client One of the adminstration consoles for SAFR. Use the Desktop Client to connect to a camera, video feed, detect faces, and submit images to the SAFR Server or Cloud for recognition.
  • macOS 10.12 or later
  • Dual Core i7
  • 1GB RAM per connected camera
  • 0.5GB available storage
  • Supports one 4K camera1
  • Supports 2+ 1080p cameras1
  • iMac Pro:
  • macOS 10.12 or later
  • 8-core Intel Xeon
  • 1GB RAM per connected camera
  • 0.5GB available storage
  • Supports up to eight 4K cameras1
  • iMac, MacBook Pro, Mac mini:
  • macOS 10.12 or later
  • 6-core Intel i7
  • 1GB RAM per connected camera
  • 0.5GB available storage
  • Supports up to three 4K cameras1
  • Supports 4+ 1080p cameras1
SAFR Actions Actions allow you to create and manage responses to event triggers; deploy them to unlock a door, turn on a light, send an alert, record data for reporting, or any security response to fit the use case.
  • macOS 10.12 or later
  • Dual Core i5
  • 1GB RAM
  • 0.5GB available storage
  • macOS 10.12 or later
  • Quad Core i5
  • 1GB RAM
  • 0.5GB available storage
SAFR Server2 The trusted engine of SAFR solutions, the SAFR Server includes: the facial recognition server, identity database, recognition event server, event archive, report server, and remote video feed administration servers.
  • macOS 10.12 or later
  • Quad Core i7 2.6GHz
  • 16GB RAM
  • 1TB available storage
  • Supports one 4K camera1
  • iMac Pro:
  • macOS 10.12 or later
  • 10-core Intel Xeon
  • 32GB RAM
  • 1TB available storage
  • Supports up to six 4K cameras1
  • Supports 7+ 1080p cameras1
  • Mac mini:
  • macOS 10.12 or later
  • 6-core Intel i7 3GHz
  • 32GB RAM
  • 1TB available storage
  • Supports up to four 4K cameras1
  • Supports 5+ 1080p cameras1

1 = Number of cameras is based on an average of five visible faces in a 4K resolution camera view, running at 15 frames per second. Using fewer faces per camera and lower resolution will enable support for more cameras.

2 = Installed as part of the SAFR Platform installer.

Linux Requirements

Product Description Minimum Requirements Recommended Requirements
Desktop Client Not available on Linux. N/A N/A
SAFR Actions Not available on Linux. N/A N/A
SAFR Server1 The trusted engine of SAFR solutions, SAFR Server includes: the facial recognition server, identity database, recognition event server, event archive, report server, and remote video feed administration servers.
  • Ubuntu 18.04(.2+), Ubuntu 20.04, Redhat 7.x, CentOS 7.x, or Amazon Linux 2018.03
  • Intel Core i5-8259U or AMD Ryzen 7 2700X
  • Quadro P2000
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • 16GB RAM
  • 1TB available storage
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
  • Ubuntu 18.04(.2+), Ubuntu 20.04, Redhat 7.x, CentOS 7.x, or Amazon Linux 2018.03
  • Intel Core i9-7980XE or AMD Ryzen TR 3700
  • Quadro RTX 5000 or Tesla T4
  • SAFR versions earlier than 3.1 are only compatible with NVidia driver versions 418.96 to 431.86
  • 32GB RAM
  • 1TB available storage
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0

1 = Installed as part of the SAFR Platform installer.

Jetson Requirements

Product Description Minimum Requirements Recommended Requirements
Desktop Client Not available on Jetson. N/A N/A
SAFR Actions Not available on Jetson. N/A N/A
SAFR Server1 The trusted engine of SAFR solutions, SAFR Server includes: the facial recognition server, identity database, recognition event server, event archive, report server, and remote video feed administration servers.
  • Ubuntu 18.04(.2+)
  • 6GB RAM
  • 5.5GB available storage
  • Jetson TX2
  • Jetson Xavier
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
  • Ubuntu 18.04(.2+)
  • 6GB RAM
  • 5.5GB available storage
  • Jetson TX2
  • Jetson Xavier
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0

1 = Installed as part of the SAFR Platform installer.

Mobile Requirements

Product Description Minimum Requirements Recommended Requirements
Mobile Client for iOS Set up a registration kiosk, perform facial recognition, and add users — all from a mobile device.
  • iOS 11.0 or later
  • iPad Pro or iPhone 6/7/8/X
  • iOS 11.0 or later
  • iPad Pro or iPhone 6/7/8/X
Mobile Client for Android Set up a registration kiosk, perform facial recognition, and add users — all from a mobile device.
  • Android 5.0 or later with Google Play Services 13.2.74 or later
  • Quad-core Snapdragon 802 2.5GHz
  • 2GB RAM
  • 13MB available storage
  • Android 6.0 or later
  • Quad-core Snapdragon 802 2.5GHz
  • Samsung Galaxy Tab S4
  • Samsung Galaxy S8
  • Google Pixel 2 XL
  • 2GB RAM
  • 13MB available storage
SAFR Beam for Android This SAFR utility allows you to configure Android mobile devices for secure SAFR operation.
  • Android 6.0 or later
  • Near-Field Communication (NFC) support required
  • 1MB RAM
  • 8MB available storage
  • Android 6.0 or later
  • Near-Field Communication (NFC) support required
  • 1MB RAM
  • 8MB available storage

SDK Requirements

Product Description Minimum Requirements Recommended Requirements
Windows SAFR SDK, Lite Edition Create a Windows app that can be used to locate and track faces and/or badges in a video file or live video stream. The Lite Edition lacks GPU acceleration, but it has a smaller footprint.
  • Windows 8.1 64-bit or later
  • C# 7.0
  • 1GB RAM per 4k video stream
  • 60MB available storage
  • Windows 10 64-bit or later
  • Microsoft Visual C++ (MSVC) 2017 or newer is strongly recommended
  • C# 7.0
  • 1GB RAM per 4k video stream
  • 60MB available storage
Windows SAFR SDK, Standard Edition Create a Windows app that can be used to locate and track faces and/or badges in a video file or live video stream. The Standard Edition has GPU acceleration. Note that the Recommended Requirements are for a single stream. For multiple streams see the Windows Desktop Client requirements.
  • Windows 8.1 64-bit or later
  • C# 7.0
  • 1GB RAM per 4k video stream
  • 0.5GB available storage
  • NVIDIA GTX 1030
  • NVIDIA driver 418.96 or later
  • Windows 10 64-bit or later
  • Microsoft Visual C++ (MSVC) 2017 or newer is strongly recommended
  • C# 7.0
  • 1GB RAM per 4k video stream
  • 0.5GB available storage
  • NVIDIA GTX 1030
  • NVIDIA driver 418.96 or later
Linux SAFR SDK, Lite Edition Create a Linux app that can be used to locate and track faces and/or badges in a video file or live video stream. The Lite Edition lacks GPU acceleration, but it has a smaller footprint than the Standard Edition.
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • You may need to downgrade the OpenSSL installation to version 3.
  • 1GB RAM per 4k video stream
  • 60MB available storage
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • You may need to downgrade the OpenSSL installation to version 3.
  • 1GB RAM per 4k video stream
  • 60MB available storage
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
Linux SAFR SDK, Standard Edition Create a Linux app that can be used to locate and track faces and/or badges in a video file or live video stream. The Standard Edition has GPU acceleration.
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • You may need to downgrade the OpenSSL installation to version 3.
  • 1GB RAM per 4k video stream
  • 0.5GB available storage
  • NVIDIA GTX 1030
  • NVIDIA driver 418.96 or later
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • You may need to downgrade the OpenSSL installation to version 3.
  • 1GB RAM per 4k video stream
  • 0.5GB available storage
  • NVIDIA GTX 1030
  • NVIDIA driver 418.96 or later
Install the following additional software components to allow VIRGO to run successfully:
  • libcurl4
  • libgomp1
  • libatomic1
  • libbsd0
  • libv4l-0
macOS SAFR SDK Create a macOS app that can be used to locate and track faces in a video file or live video stream.
  • macOS 10.12 or later
  • 1GB RAM per 4K video stream
  • 215MB available storage
  • macOS 10.14 or later
  • 1GB RAM per 4K video stream
  • 215MB available storage
iOS SAFR SDK Create an iOS app that can be used to locate and track faces in a video file or live video stream.
  • iOS 11 or later
  • iPhone 6
  • Swift 5
  • 92MB available storage
  • iOS 12 or later
  • iPhone X or iPad Pro
  • Swift 5
  • 92MB available storage
Android SAFR SDK Create an Android app that can be used to locate and track faces in a video file or live video stream.
  • Android 6.0 or later
  • 1GB RAM
  • 0.5GB available storage
  • Android 6.0 or later
  • 1GB RAM
  • 0.5GB available storage

Embedded SDK Requirements

Product Description Minimum Requirements Recommended Requirements
Windows x86 SAFR Embedded SDK, Lite Edition Build a facial recognition app on a Windows device with limited resources (RAM, CPU, or memory). The Lite Edition lacks GPU acceleration, but it has a smaller footprint than the Standard Edition.
  • Windows 8.1 64-bit or later
  • x86 Architecture
  • 200MB RAM
  • 60MB available storage
  • Windows 10 64-bit or later
  • x86 Architecture
  • 200MB RAM
  • 60MB available storage
Windows x86 SAFR Embedded SDK, Standard Edition Build a facial recognition app on a Windows device with limited resources (RAM, CPU, or memory). The Standard Edition has GPU acceleration.
  • Windows 8.1 64-bit or later
  • x86 Architecture
  • 200MB RAM
  • 0.5GB available storage
  • NVIDIA GTX 1030 or better
  • NVIDIA driver 418.96 or later
  • Windows 10 64-bit or later
  • x86 Architecture
  • 200MB RAM
  • 0.5GB available storage
  • NVIDIA GTX 1080 Ti
  • NVIDIA driver 418.96 or later
Linux x86 SAFR Embedded SDK, Lite Edition Build a facial recognition app on a Linux x86 device with limited resources (i.e. RAM, CPU, or memory). The Lite Edition lacks GPU acceleration, but it has a smaller footprint than the Standard Edition.
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • x86 Architecture
  • 500 MB RAM
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • x86 Architecture
  • 500 MB RAM
Linux x86 SAFR Embedded SDK, Standard Edition Build a facial recognition app on a Linux x86 device with limited resources (RAM, CPU, or memory). The Standard Edition has GPU acceleration.
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • x86 Architecture
  • 1500 MB RAM
  • Nvidia GPU GTX10xx or later
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • x86 Architecture
  • 1500 MB RAM
  • Nvidia GPU GTX10xx or later
Linux ARM SAFR Embedded SDK Build a facial recognition app on a Linux ARM device with limited resources (RAM, CPU, or memory).
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • 64bit ARMv8 CPU
  • 200 MB RAM
  • Ubuntu 18.04(.2+) 64-bit or Ubuntu 20.04
  • 64bit ARMv8 CPU
  • 200 MB RAM
Jetson SAFR Embedded SDK Build a facial recognition app on a Jetson device with limited resources (RAM, CPU, or memory). The following Jetson devices are supported:
  • Nvidia Jetson TX2
  • Nvidia Jetson Xavier
  • Nvidia Jetson Nano
The following Jetson devices are supported:
  • Nvidia Jetson TX2
  • Nvidia Jetson Xavier
  • Nvidia Jetson Nano
Android ARM SAFR Embedded SDK Build a facial recognition app on an Android device with limited resources (RAM, CPU, or memory).
  • Android 6.0 or later
  • ARMv7 or ARVMv8 Architecture
  • 200MB RAM
  • 150MB available storage
  • Android 6.0 or later
  • ARMv7 or ARVMv8 Architecture
  • 200MB RAM
  • 150MB available storage