Twelve Labs
Overview
Multimodal AI platform for video understanding — search, analyze, classify, and embed video content using advanced foundation models like Marengo and Pegasus.
Full Description
Twelve Labs is a multimodal AI platform that enables developers to build applications that can see, listen to, and understand video content at a human-like level. Powered by foundation models Marengo (video encoder) and Pegasus (video-language model), the platform provides natural language video search, automated analysis and summarization, zero-shot classification, and vector embeddings for recommendation systems. It processes visual frames, speech, on-screen text, and audio simultaneously. Available on AWS Bedrock, Databricks, and Snowflake with Python and JavaScript SDKs. Backed by NVIDIA, Samsung, Intel, and Databricks Ventures, the platform serves over 30,000 developers across media, security, advertising, and content platforms.