Amazon Bedrock Adds Marengo 3.0 for Semantic Video Search

It is a multimodal embedding model from TwelveLabs that converts video, image and audio into vectors for natural language search, now generally available inside Amazon Bedrock Knowledge Bases.

Traditional search relies on metadata or extracted text, while Marengo 3.0 understands the visual and audio content itself, allowing queries like 'cars in night rain' without pre-tagging.

Yes, particularly in media, government archiving, security and training, where large unindexed video libraries exist. Regional cloud regions can reduce latency and support compliance.

If you have a large media library and need semantic search, piloting via Bedrock is low risk. Regulated entities should first confirm data processing location and compliance.