Moments Lab, an AI video discovery company, is set to unveil MXT-2 at the 2025 NAB Show. MXT-2 is the latest iteration of their award-winning AI indexing technology. This multimodal and generative AI solution provides human-like descriptions of video content, significantly reducing the time spent searching through large video libraries.
MXT-2 leverages AI-generated, time-coded metadata to enhance video search capabilities. It can identify specific people, places, and logos, and detects relevant moments based on content type. This improved version boasts a larger, more sophisticated AI model trained on meticulously curated data, resulting in more accurate and natural-sounding descriptions. The technology is accessible through the Moments Lab platform or its public API.
“We’ve designed MXT-2 to provide organizations with deeper editorial insights and video searchability at scale,” said Philippe Petitpont, co-founder and CEO at Moments Lab. “At NAB this year, we’re particularly excited to unveil the latest features added to our patented AI indexing technology, which is built using RAG (retrieval-augmented generation). RAG maximizes content creation, reuse, and repurposing, delivering tangible ROI by enabling entire media libraries to be described in natural language and therefore searched in this same language.”
Moments Lab will offer demonstrations of its solutions at the 2025 NAB Show. In addition, Khalil Guetari, AI software engineer at Moments Lab, will present "Vision and Language Models for Enhanced Archive Video Management" at the BEIT Conference (April 5-8). This session is scheduled for Monday, April 7, 10:40 AM – 11:00 AM PDT in Room W222-W223.