Eluvio Debuts Inline AI and Agentic Orchestration for Video Pipelines

The Berkeley-based company’s new Universal & Dynamic Video Intelligence architecture functions natively within the Content Fabric, allowing AI models to execute during streaming. This approach replaces conventional cloud workflows that typically require copying data into separate processing environments. By keeping the entire process in place, Eluvio achieves zero-copy, zero-movement analysis, which significantly reduces latency and infrastructure costs for broadcasters and rights holders.

At the core of this release are 17 built-in models and a new Model Context Protocol (MCP) API. This framework allows external agents, such as LLMs, to perform complex, natural-language-driven tasks like automatic highlight generation, scene segmentation, and dynamic vertical video cropping. CEO Michelle Munson noted that the system is designed to provide immediate actionability, from identifying specific players in a live match to generating personalized summaries. These capabilities, alongside the updated Eluvio Video Intelligence Editor (EVIE), will be demonstrated at the RAI Amsterdam from September 11 to 14, 2026.