Scrypted Docs

Appearance

Features ​

Smart Detections ​

Timeline and Events ​

The Timeline and Events views are powered by Smart Detections that highlight key moments in the day: people and cars arriving and leaving, packages, and animals.

Timeline
Events

Face Recognition ​

Tag and recognize faces for use in automations.

License Place Recognition ​

Recognize and search vehicle plates.

Rich Notifications ​

Rich Notifications are delivered instantly to iOS, Android, and Web apps with a thumbnail of the detection.

Search across the history of all cameras using natural language.

AI Summaries ​

Scrypted NVR can use AI to automatically generate natural-language descriptions of camera events. When detections occur, the AI produces a short summary and title that is delivered as a push notification and displayed in the NVR app's Stories tab. Setup guide.

Adaptive Bitrate ​

Scrypted NVR leverages all the camera substreams and transcoding to provide bulletproof streams, regardless of downstream network conditions.

TIP

Adaptive Bitrate is a Platform level feature, and also improves streaming to HomeKit, Google Home, Alexa, and Home Assistant.

Hardware Accelerated AI ​

Scrypted NVR Smart Detections use one of the available Object Detection Plugins which can run on:

  • CPU
  • GPU
  • Apple Silicon Neural Engine 🔥
  • Intel Core Ultra NPU 🔥
  • Coral.ai accelerators

Apple Silicon ​

Scrypted NVR can leverage the Neural Engine on the new Apple Silicon chips using the CoreML Plugin. Incredible performance at low power. Installing the Scrypted Desktop App is recommended on Mac.

Intel ​

Scrypted NVR supports Intel iGPUs and accelerator cards for blazing fast detection performance using the OpenVINO Plugin. 12th Generation Intel processors like the N100 have great performance at a low power and price point.

Intel Core Ultra chips with the built in NPU are also supported and provide state of the art performance.

NVIDIA ​

Scrypted NVR supports object detection on NVIDIA GPUs via the ONNX Plugin. NVIDIA object detection requires CUDA, which is only provided in Scrypted's NVIDIA Docker image when using the koush/scrypted:nvidia image.

Google Coral EdgeTPU ​

Scrypted NVR supports object detection using the Google Coral.ai accelerators via the Tensorflow-Lite Plugin. The PCI and USB models are both supported.