Twelve Labs

Twelve Labs is an artificial intelligence (AI) start-up developing AI models and tools for understanding videos. Twelve Labs' technology can extract features from video, including actions, objects, text on screen, speech, and people. This information is transformed into vector representations for semantic search. The company's AI models provide context-specific search and insights, replacing keyword tagging to allow users to search their videos. The models can be fine-tuned by the user to meet domain-specific needs.

Twelve Labs AI models can be accessed via a suite of APIs, engines API for selecting Twelve Labs default engine or the user's fine-tuned model, index API for uploading and vectorizing videos, task API (search and vector API), dataset API providing custom labels for input videos, finetune API for customizing the default model, and deploy API for adding fine-tuned models to future selections.

Use cases for Twelve Labs' technology include the following:

• Contextual advertising
• Content moderation
• Evidence search
• Content search
• Media analytics
• Digital asset management
• Lecture search
• Video recommendation
• Video editing

Twelve Labs was founded in 2021 by Jae Lee (CEO), Aiden Lee (CTO), SJ Kim (head of engineering), Dave Chung (head of operations), and Soyoung Lee (head of business development). The company took part in Techstars Seattle 2021 class. The company has raised $27 million in seed funding and is backed by Index Ventures, Radical Ventures, Expa, Wndrco, and a number of leaders in AI, including Fei-Fei Li, Silvio Savarese, Oren Etzioni, Minjoon Seo, Aidan Gomez, Nick Frosst, and Ivan Zhang. On October 23, 2023, Twelve Labs introduced Pegasus-1, a video-language foundation model as well as a new suite of video-to-text APIs (Gist API, Summary API, Generate API).