The Microsoft Cognitive Services Video API provides cloud based face detection and tracking, and supports detecting motion, stabilizing, and creating thumbnails from video. This API can detect up to 64 human faces; frontal faces provide better results. Additionally, it can be used to stabilize surveillance and shaky videos, and to generate a preview or snapshot of a video. Currently, a free plan that limits transactions to 300 per month per feature is available. The Video API uses JSON format for data exchange, and API Keys for authentication.