Eluvio is set to unveil a major new architecture for universal and dynamic video intelligence and next-gen Eluvio Video Intelligence Editor (EVIE) with advanced AI tools for agentic orchestration of title libraries and live sports ahead of NAB Show 2026.
Eluvio AI is said to be the first commercially available solution that runs AI analysis and inference inline within the streaming media generation and distribution pipeline, producing frame accurate, fully-aligned textual and multi-modal labels, embeddings and metadata, for both live and VoD content. It harnesses AI data just-in-time, enabling unlimited AI personalisation and transformation of video, audio and image content.
This first-of-its-kind implementation operates with zero file copies, zero file movement, and zero re-transcoding at every stage: analysis, metadata generation, derivative creation, and delivery. It is a new architectural foundation for AI usage in live sports, events, studio archives, and other premium video use cases worldwide.
Key benefits for users are personalisation and re-monetisation enabled by multi-modal, frame-accurate deep analysis and search (video, audio, text, images, and pose), enhanced efficiency compared to other video AI workflows with zero file copies, file movement, or re-transcoding, and unlimited runtime API and vector/tag stores (with 15+ built-in models and processors) and fully open for continuous addition of the state-of-the-art without changing workflows or moving media.
Additional features include fully private ground truth, training, model configuration, and content protection with Content Fabric owner-controlled security and fine-grained permissions; unlimited generative power for creating high quality vertical video from 16:9, and suggesting highlights, shorts and more from content using the zero-copy JIT features of the Content Fabric; and multi-agent orchestration APIs for bringing it all together behind natural language prompts, third-party chatbots, and agentic interfaces.
“What broadcasters, sports leagues, and entertainment companies need from AI right now is operational reality and the right architecture to flex continuously with the extreme pace of change — frame accuracy across every modality – from low bitrate audio to high frame rate video and multi-D telemetry, zero-copy efficiency, live inference, and the ability to generate unlimited derivative content without duplicating assets or moving files,” said Michelle Munson, CEO and co-founder of Eluvio. “Our architecture delivers that, natively, inside the Eluvio Content Fabric. EVIE and our orchestration APIs then make it all accessible, and easy-to-use, for any media professional: a single pipeline from live stream to progressive VOD, with in-place multi-AI assistance for personalized highlights, topic-based channels, and infinite re-combinations. We have proven this in partnership with our customers, leading sports and entertainment organisations on a global scale, and look forward to demonstrating what is possible at NAB 2026.”
The newest Eluvio Video Intelligence Editor (EVIE) is an AI-native editing and workflow application, enabling infinite re-combinations: personalised highlights, topic-based compositions, and AI-assisted zero-copy derivative content created in-place.
The new EVIE Titles & Events (Beta preview) feature unifies the full AI stack of tools under a single UX and agentic API indexed by title (film/TV) or event (live sports). Capabilities for media professionals, include unlimited deep content search across the library; automatic personalised game highlights and chapter segmentation combining play-by-play, ball and player tracking, audio analysis and key moment identification; audio transcription; motion & scene analysis; topic, mood & theme detection; automatic summaries & captions, and tailored synopsis for marketing, sales and social; and the ability to create vertical video from any 16:9. EVIE automatically detects focal points by shot and class, generates per-shot crop coordinates, and enables the Fabric to produce 9:16 vertical video from 16:9, just-in-time — dynamically, without re-transcoding or copying the source, and applicable to live streams.
