Google Cloud Video AI

Video AI, built on the Video Intelligence API, analyzes stored and streaming video to recognize more than 20,000 objects, places and actions. Media companies and developers use the metadata it returns to search catalogs, filter content and build features such as highlight reels.

Annotations can be returned at the level of a whole video, a shot or a single frame. Features include label detection, shot change detection, explicit content detection, speech transcription, object tracking, and text, logo, face and person detection.

Streaming annotation gives near real-time results and can trigger events when a chosen object appears. Billing is per minute of video with partial minutes rounded up, the first 1,000 minutes of each feature per month are free, and customers above 100,000 minutes a month can ask for discounts.