TwelveLabs video understanding models now available in Amazon Bedrock

Amazon Web Services added TwelveLabs as a new model provider to Amazon Bedrock, delivering what could be the most significant breakthrough in enterprise content management — making the world’s vast video archives instantly searchable.

Video contains nearly 80% of the world’s data, yet most remains completely unsearchable. TwelveLabs’ multimodal foundation models (FMs) use advanced video understanding to interpret what is happening in videos, including actions, objects, and background sounds. Their state-of-the-art models are now available through Amazon Bedrock’s model marketplace. AWS becomes the first cloud provider to offer this technology to customers.

• Marengo 2.7: Get fast, context-aware results that reveal exactly what you’re looking for in your videos — moving beyond basic tags into a whole new dimension of multimodal understanding.
• Pegasus 1.2: Transform your videos through the power of language. Generate everything you need from video — from concise summaries and captivating highlights to effective hashtags and customized reports. Uncover deeper insights and unlock entirely new possibilities with your content.

Real-World Impact:

• Film/TV Studios: Rapidly manage dailies, content repacking, and archive management
• Streaming Services: Better content packaging and targeted ad insertion
• Sports Networks: Instantly pull every instance of specific play styles or commentator reactions
• News Agencies: Find “moments that matter” across vast libraries