TwelveLabs' Pegasus 1.5 now available on Amazon Bedrock with two-hour video analysis

TwelveLabs' Pegasus 1.5 video language model is now available on Amazon Bedrock, enabling users to analyze videos up to two hours long in a single request. The model returns structured, timestamped JSON metadata for every segment matching user-defined criteria, supporting multimodal prompts that combine text and reference images to locate specific products, logos, or people. Processing occurs on-the-fly directly from Amazon S3 URIs or base64 strings without pre-indexing, and video data remains within the AWS data boundary.