TwelveLabs is an AI-powered video intelligence platform that enables developers and enterprises to understand, search, and analyze video content at scale. Built on two foundation models — Marengo (a multimodal video embedding model) and Pegasus (a video-first language model) — TwelveLabs provides APIs for semantic video search, automatic content analysis, and multimodal embedding. Pegasus 1.5, unveiled at NAB Show 2026, introduces time-based metadata extraction and temporal reasoning across videos up to two hours long, outperforming Gemini 2.5 Pro by 30% on aggregate segmentation benchmarks. The platform supports use cases including automatic video tagging, scene detection, chapter generation, content summarization, contextual ad matching, and natural language video search across speech, text, audio, and visual dimensions. TwelveLabs also launched Rodeo, an application-layer creative co-pilot that lets editors find, edit, and assemble footage using natural language. Developers can integrate via Python and Node.js SDKs, and enterprise deployments are supported with custom SLAs.