Inline, Open-Model Video AI and Agentic Orchestration at IBC 2026

• Universal & Dynamic Video Intelligence Architecture: First commercially available solution for inline, frame-accurate, multimodal AI inference, analysis, and orchestration built natively into the Content Fabric media pipeline for live and derivative content generation
• New Open Model AI Inference Engine with Fast Text & Vector Storage: Executes AI inline with content distribution and streaming, allowing any frame or video-level model to execute directly in the Fabric distribution pipeline, with inference metadata instantly available globally to the Fabric’s just-in-time media pipeline
• 17 Built-In Models and Extensible API: Covers analysis, segmentation, summarization, and multi-language transcription, with an extensible API for third-party and custom models. Unlike hypervisor cloud solutions, video AI inference requires no media or metadata movement, and AI data can directly drive frame-accurate personalization of live and on-demand content
• Newest Models and Agentic Processors: Include just-in-time live vertical video generation combining live sports motion analysis, expert focus tracking, and per-player jersey identification; expanded shot, beat, and scene segmentation; “In this video” key moment identification; and personalized synopses and summaries by segment or across entire titles and events
• Next-Generation Eluvio Video Intelligence Editor (EVIE): Expanded with EVIE Titles, a unified title-view of AI intelligence; a new AI Runtime for inference management; media-synchronized visualization of AI data by frame, shot, beat, and scene; enhanced ground-truth management; zero-copy AI clipping and composition; automatic dynamic vertical video playout; and a new Eluvio MCP API exposing major functions for agentic orchestration
• Eluvio at IBC Show 2026: Hall 8, Stand 8.MS5, and the Future Tech Zone, Hall 14, Stand 14.A58, September 11–14, 2026, RAI Amsterdam

BERKELEY, Calif., Sept. 3, 2026 /PRNewswire/ — Eluvio today announced that it will showcase major new advances in its Universal & Dynamic Video Intelligence architecture and Eluvio Video Intelligence Editor (EVIE) at IBC Show 2026, introducing an expanded open-model AI inference engine, 17 built-in AI models and processors, fast text and vector storage, new live sports intelligence and vertical video capabilities, and Model Context Protocol (MCP) API for agentic orchestration across live events and title libraries.

Eluvio AI is the first commercially available architecture to run frame-accurate, multimodal AI analysis and inference inline within the streaming media generation and distribution pipeline itself. AI models operate directly against live and on-demand video as it flows through the Content Fabric, producing fully time-aligned labels, embeddings, vectors, transcripts, and metadata that are immediately available to the Fabric’s just-in-time media pipeline.

Unlike conventional cloud AI workflows that require video to be copied, moved into separate AI processing infrastructure, analyzed, stored again, and frequently re-transcoded to produce derivatives, Eluvio performs the entire workflow in place. There are zero file copies, zero media or metadata movement, and zero re-transcoding across analysis, metadata generation, personalization, derivative creation, and delivery.

The result is a single architecture in which AI intelligence can immediately drive the video itself — enabling frame-accurate search, personalization, clipping, highlights, vertical video, compliance reels, summaries, and other derivative experiences for both live and on-demand content. Eluvio will demonstrate these capabilities live at IBC Show 2026.

“Eluvio AI is fundamentally different from point AI solutions because the intelligence runs inline, with frame accuracy, across live and archival content — without the cost and delay of copying and re-transcoding media before it can be processed and ensembling the power of individual models with frame-accuracy,” said Michelle Munson, CEO and co-founder of Eluvio. “The AI is built directly into the video pipeline of the Content Fabric, rather than added afterward as a separate cloud workflow. Everything from identifying action, players, celebrities, speakers, or brands to turning that intelligence into highlights, vertical videos, compliance reels, or personalized content happens in place against the source media and has no limits on size, duration, or ensemble scale. The result is dramatically greater efficiency and speed, while making AI intelligence immediately actionable for personalization and monetization. At IBC, we’re excited to show broadcasters and rights holders these capabilities running across both live sports and premium content.”

Universal & Dynamic Video Intelligence Architecture

At the foundation of the new capabilities is Eluvio’s Universal & Dynamic Video Intelligence Architecture, which integrates AI inference, multimodal intelligence, text and vector storage, search, generation, and agentic orchestration directly into the Content Fabric media pipeline. Its key capabilities include:

• Frame-Accurate & Multimodal Intelligence — Video, Audio, Text, Images, Pose & Telemetry: Inference operates at frame and segment level across video, audio, text, images, pose, and multidimensional telemetry. Because the intelligence is generated against the underlying media timeline, no secondary metadata alignment step is required.
• Zero-Copy, Zero-Movement AI: Models execute directly against Content Fabric media parts. Media does not need to be exported into separate AI pipelines, and resulting intelligence is written directly back into the Fabric where it is immediately available for search, personalization, generation, and distribution.
• Real-Time Inline Inference for Live and VoD: AI reads live content parts as they are published and writes frame-accurate intelligence back in real time, enabling the same architecture to operate across live sports, live broadcasts, and on-demand libraries.
• AI-Driven Just-in-Time Media: AI metadata becomes directly actionable by the Content Fabric’s just-in-time media pipeline, allowing intelligence generated at a specific frame, shot, beat, scene, player, speaker, or event to dynamically control clipping, composition, personalization, transformation, and delivery.

New Open Model AI Inference Engine, 17 Built-In Models & Fast Text/Vector Storage

Eluvio’s expanded AI inference engine provides an open runtime for frame-level, video-level, multimodal, and language models, allowing new open-source, Eluvio-developed, customer-specific, and third-party models to be incorporated without changing the underlying media workflow or moving the media. The platform now includes 17 built-in models and processors spanning multimodal analysis, segmentation, summarization, transcription, identification, search, and generation. Capabilities include:

• Frame and Video-Level Inference: Models can operate against intra-coded frames, randomly sampled frames, segments, shots, scenes, or entire videos.
• Shot, Beat & Scene Segmentation: Expanded temporal analysis identifies meaningful structural boundaries within content, providing the foundation for search, summarization, clipping, and composition.
• Multimodal Embeddings: OpenCLIP, ImageBind, Eluvio-enhanced MediaPipe, and other models support image, video, audio, text, and pose intelligence.
• Motion, Focus & Identity Analysis: New capabilities include custom motion identification, expert focus tracking, speaker detection and identification, celebrity identification, and player identification.
• Multi-Language Transcription: Speech-to-text produces time-aligned transcripts that can be combined with speaker recognition and other visual intelligence.
• Fast Text, Tag & Vector Storage: Model outputs, tags, embeddings, and associated metadata are stored directly within the Fabric and made globally available for immediate search, retrieval, inference, personalization, and generation.
• Extensible Runtime API: Custom, open-source, and third-party models can be added continuously as the state of the art evolves, without redesigning workflows or creating separate copies of the underlying media.
• Private Ground Truth & Model Configuration: Customers can create and manage proprietary ground-truth datasets and deploy custom models trained against their own labeled content within the Content Fabric’s tenant-isolated security model.

Newest Models & Agentic Processors for Live Sports and Premium Video

At IBC, Eluvio will demonstrate a new generation of models and processors that turn AI analysis directly into actionable media outputs. For live sports, these capabilities include just-in-time vertical video generation combining live motion analysis, expert focus tracking, and per-player jersey identification. The system dynamically determines the focal point for each shot and produces 9:16 vertical presentations from a live 16:9 source — without creating a second encoded video stream or pre-generating separate vertical assets. Additional new processors include:

• “In This Video” Key Moment Identification: Automatically identifies the most important moments within a title, event, segment, or scene.
• Personalized Synopsis & Summarization: Generates configurable synopses and summaries for individual segments or complete titles and events, tailored by purpose, length, and style.
• Live and On-Demand Highlights: Combines AI analysis with play-by-play, player and ball tracking, audio intelligence, and existing metadata to identify and compose personalized highlights.
• Automatic Chapters & Segmentation: Generates chapter boundaries, labels, and descriptions using shot, beat, scene, topic, mood, and semantic analysis.
• Zero-Copy Generative Composition: Highlights, trailers, shorts, social clips, compliance reels, and other compositions are represented as references to the original media rather than newly rendered files.
• Natural-Language Personalization: Agentic processors can use natural-language prompts to control the purpose, content, length, style, and audience of generated outputs.

Next-Generation EVIE and EVIE Titles & Events

The Eluvio Video Intelligence Editor (EVIE) provides the media-facing application layer for this intelligence, allowing users to inspect, search, validate, edit, and act upon AI data while remaining synchronized to the underlying video. The new EVIE Titles & Events experience brings the complete AI stack together around a title — such as a film or television program — or an event, such as a live sports match. New and expanded capabilities include:

• AI Runtime: A new inference-management environment for selecting, running, and managing AI models and processors against content.
• Media-Synchronized AI Visualization: Model outputs are organized as synchronized tracks against the media timeline and can be viewed at frame, shot, beat, scene, segment, title, or event level.
• Deep Multimodal Search: Search across entire content libraries using natural language, images, text, music, metadata, and visual regions to identify matching shots, scenes, clips, or individual frames.
• Ground-Truth Management: Expanded tools allow media professionals to review and correct AI results and manage private ground truth, including celebrity, speaker, and other identity data.
• Analyze: Combine transcription, speaker recognition, motion and scene analysis, topics, moods, themes, pose, telemetry, celebrities, brands, and other intelligence within a single synchronized environment.
• Highlight & Compose: Generate personalized sports highlights, chapter reels, trailers, shorts, compliance reels, and social compositions without duplicating the underlying source media.
• Dynamic Vertical Video: Automatically identify focal points and crop coordinates shot-by-shot and generate dynamic 9:16 video directly from 16:9 live or on-demand sources through the Fabric’s just-in-time pipeline.

New MCP API for Agentic Video Orchestration

Eluvio is also introducing a Model Context Protocol (MCP) API exposing major Content Fabric and Eluvio AI functions for agentic orchestration, allowing ChatGPT, Claude, other LLMs, third-party applications, and agentic systems to execute complex video workflows through natural-language interfaces.

A single agentic workflow can invoke models, analyze content, update vector indices, search video, generate transcriptions, identify key moments, create chapters and highlights, and produce social-ready clips, trailers, shorts, and other compositions. Because the underlying APIs operate directly against Fabric content, these agentic workflows retain the same zero-copy architecture and Content Fabric authorization model. Media remains secured within its tenant and owner-controlled permissions while agents operate against authorized functions and intelligence.

Separately, Eluvio today announced full details of the Global Broadcast Fabric, the newest Content Fabric version spanning broadcast distribution, D2C OTT streaming and monetization, and inline AI video intelligence in a companion press release 

Eluvio at IBC Show 2026

During IBC Show 2026 at the RAI Amsterdam, Eluvio will demonstrate Eluvio AI, EVIE, live sports intelligence, agentic video workflows, and the full capabilities of the Content Fabric in Hall 8, Stand 8.MS5, and in the Future Tech Zone, Hall 14, Stand 14.A58, September 11–14, 2026.

On Thursday, September 10, Eluvio execs will be present at Eluvio’s table at the SVGE Europe Summit at De Hallen Studio. On Friday, September 11 from 7:30-10:30 am, Michelle Munson will speak on the “Where does the Signal Go Next” broadcast transition executive breakfast at the Traveler within the RAI. On Saturday, September 12 at 1:00 pm, Eluvio founders Michelle Munson and Serban Simu will deliver a keynote deep dive at the Eluvio stand. On Sunday, September 13 from 2pm-2:15 pm Michelle Munson will present “Beyond Cloud and Satellite: The New Architecture for Broadcast” on the Future Tech Zone stage in Hall 14.

Registration for one-on-one meetings and information about Eluvio’s on-site keynote demonstrations at IBC is available at ibc.eluv.io.

About Eluvio

Eluvio is solving the generational problems of video over the Internet with its Content Fabric Protocol — transforming the distribution and monetization of premium video and digital media in contrast to legacy cloud workflows and CDNs. Eluvio serves Amazon Studios/MGM, Cricket Australia, European Professional Club Rugby, SONY, Telstra Broadcast Services, UEFA, United Rugby Championship, Warner Bros., WPALiveTV, WWE, and others.Led by Michelle Munson and Serban Simu, founders and inventors of Aspera, Eluvio is based in Berkeley, California. Eluvio holds 15 issued U.S. patents and has won multiple NAB Streaming Product of the Year awards, the Seicon Grand Prize, the Hollywood Professional Association Engineering Excellence Award, and other industry honors.

@EluvioInc | linkedin.com/company/eluv-io 

SOURCE Eluvio, Inc.