Original Coverage & Source Attribution: www.mk.co.kr
An artificial intelligence (AI) secretary who finds the scene he saw on YouTube well on his own and neatly came out. It can be easily used by ordinary users without having to look for it again while watching a long video.
Twelb Labs announced on the 16th that it has officially launched an AI agent called “Jockey by TwelveLabs” that provides such services.
Twelve Labs is a multimodal video understanding AI company. It develops a general-purpose video understanding technology that understands the behavior, events, and context of a vast amount of video in various markets such as media, entertainment, sports, and public.
“Jockey” released by Twelve Labs is an AI agent that finds related scenes in the entire video library when a user requests them in natural language and explains why the scene is important.
Based on Twelbraps’ accumulated image search and automatic metadata generation capabilities and image understanding technology to summarize images and pick up highlight sections, it quickly executes tasks requested by users in natural language and delivers results.
It is possible for Zaki to first select videos that are found beyond video search, such as views and likes, or to make an edited version of the video that is attached by selecting only the necessary scenes from the original video.
In the past, video content was not easily intimidated, unlike text. The video is not searched like text, so you have to turn it yourself or lean on additional information that only indirectly contains the video content, such as titles, tags, and subtitles, to find the scene. In other words, Zaki has solved this problem.
In particular, the jockey can be easily used by ordinary users without development knowledge. You can handle the video as if you were talking to a jockey on OpenAI’s ChatGPT or Antropic’s Claude app.
Jockey can analyze trends that are now receiving responses by entering links to social videos such as YouTube, TikTok, and Instagram Reels, and bring in participation indicators for each video to examine content performance at once.
Results such as edited or analyzed data created by Zaki can be stored together with the video source and shared with others.
With the launch of Zaki, Twelve Labs has AI competitiveness that ranges from foundation models that understand videos to agents that perform actual tasks.
“As Twelbraps’ AI capabilities accumulated so far are implemented as one integrated video understanding system, video will be reborn as a driving force for innovation that creates new value beyond just usable content assets,” said Lee Jae-sung, CEO of Twelbraps.
