Twelve Labs

Founded in 2020, Twelve Labs provides enterprises and developers with a multimodal video base model in the video domain through an API, which is mainly used for video multimodal search, where users can search for any content in a video by inputting text and pictures (text / picture to any); it also introduces the functions of video intelligent Q&A and intelligent categorization.

Twelve Labs' vision is to become ChatGPT for video, which is currently the best product in the field of video multimodal search. Its biggest highlight is that the video search effect is very accurate, able to understand abstract concepts, and is in the absolute leading position among its kind, and its customers generally evaluate its search quality is very good, search speed is fast, and generalizability is good. The barriers of the video base model are very high, from the high-quality video data, the infra that processes the data, the index system, the training method, and even to the cooperation with the chip company (investor), Twelve Labs has built a certain "first-mover advantage".