Video and image search in Amazon Bedrock Knowledge Base using Marengo 3.0
aws ·
AWS announced the general availability of TwelveLabs Marengo Embed 3.0 as an embedding model in Amazon Bedrock Knowledge Bases. Marengo Embed 3.0 is a multimodal embedding model that jointly encodes video, audio, images, and text into a 512-dimensional vector space. Amazon Bedrock Knowledge Bases is a fully managed RAG service that handles storage, ingestion, embedding, re-ranking, and retrieval, and it supports MP4 and MOV video files, JPEG and PNG images, and audio tracks. The service includes native connectors for Amazon S3, SharePoint, and Confluence, and Managed Knowledge Bases automatically generate multimodal embeddings capturing visual, textual, speech, and audio signals. The article provides a walkthrough using a 10-minute clip of the 2022 FIFA World Cup final, covering uploading media to Amazon S3, creating a managed knowledge base, selecting Marengo Embed 3.0, syncing and ingesting data, and running natural language queries such as “show me the penalty kicks from this soccer match.” Query results return ranked chunks with metadata including start time, end time, source URI, and embedding type, and applications can use the Amazon Bedrock Retrieve API for downstream integration. The offering is available in US East (N. Virginia) and US West (N. California), with customers paying only for what they store and retrieve, while embeddings generation is charged at the standard Amazon Bedrock model invocation rate.
aws