Twelve Labs Launches AI Agent Video Understanding MCP Server

Twelve Labs opens new horizons in AI! Hyperscale AI company Twelve Labs has officially launched its 'Model Context Protocol (MCP) server,' which endows AI agents with 'the ability to understand video.' Now, AI assistants can not only understand text or images but also grasp the context of moving video content, identify key scenes, and generate summaries, heralding an amazing new era. Based on MCP, an open standard protocol developed by Anthropic, the Twelve Labs MCP server connects its powerful video understanding models with major AI tools favored by developers, such as Claude Desktop and Cursor. Developers can now 'plug and play' Twelve Labs' unique video intelligence capabilities into their AI applications with ease. At the core of this server are Twelve Labs' self-developed multimodal video understanding model, 'Marengo,' and the video language generation model, 'Pegasus.' These enable a wide array of functions, including natural language-based video search, automatic video content summarization and Q&A, the construction of complex multi-step video workflows, and real-time video exploration support. From smart assistants that understand recorded meeting videos to creative content generation AI utilizing video context, the development of infinite multimodal applications has become possible. Jae-sung Lee, CEO of Twelve Labs, emphasized, "This MCP launch is a turning point where video establishes itself as a core feature in all AI workflows." He further stated, "Our philosophy has been the implementation of true 'multimodality' — not a 'multi-model' approach using separate models for text, image, and audio, but rather a unified understanding of all video elements through a single interface. The MCP server is the result of that conviction." Since its founding in 2021, Twelve Labs has rapidly grown in the global AI market, demonstrating unparalleled technological prowess. It has been recognized as one of CB Insights' 'Global 100 AI Startups' for four consecutive years, and in April, achieved the remarkable feat of being selected as the first Korean AI model provider for Amazon Bedrock. CTO Seung-jun Lee was recognized for his innovation by being named an AI leader in Forbes North America's '30 Under 30,' and the company successfully raised approximately $107 million in cumulative investment. By participating in major international events such as IBC 2025 and the Busan International Film Festival, Twelve Labs is accelerating its global market expansion, and its endeavors in leading the future of video AI are garnering further attention.