TwelveLabs has launched its Pegasus 1.5 video language model on Amazon Bedrock, allowing users to generate structured, timestamped metadata from long-form video content. The model supports multimodal prompting and on-the-fly processing, enabling automated content cataloging and highlight production without pre-indexing.
The availability of this model on Amazon Bedrock provides streaming platforms with a scalable method to convert unstructured video data into searchable JSON metadata. By automating the identification of editorial segments and brand appearances, organizations can significantly reduce the manual labor required for archive management and ad placement. This integration keeps sensitive video data within the AWS boundary, addressing security concerns for major broadcasters. As the ecosystem shifts toward hyper-personalized discovery, watch for how quickly streaming services adopt these automated workflows to generate scene-level recommendations and contextual advertising triggers.
TwelveLabs has launched its Pegasus 1.5 video language model on Amazon Bedrock, allowing users to generate structured, timestamped metadata for videos up to two hours long. This integration enables automated content cataloging and highlight production, helping streaming platforms reduce manual labor while keeping sensitive video data secure within the AWS boundary.
Pegasus 1.5 can process videos up to two hours long in a single request.
No, the model uses on-the-fly processing to analyze videos directly from Amazon S3 URIs, eliminating the need for ingestion pipelines.
The model supports 12 languages, including English, Arabic, Chinese, and Spanish.
Multimodal prompting allows users to combine text descriptions with visual reference images to locate specific products or logos within a video.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source