Twelve Labs

Multimodal AI platform providing APIs for instant, natural language search and deep understanding across petabytes of video data.

Twelve Labs delivers video-native AI models (Marengo and Pegasus) to unlock the full value of video archives for enterprises and developers. The core technology, including the Pegasus-1 video-language model, offers powerful APIs for search, summarization, and analysis, allowing users to pinpoint exact moments in massive video libraries using natural language queries. This capability drastically reduces content workflows: one customer, MLSE, cut their content ingestion time from 16 hours to just 9 minutes. We handle petabytes of data, providing the essential building blocks (embeddings, video-to-text generation) for next-generation video intelligence applications.

https://www.twelvelabs.io/