Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

FRAMES

FRAMES (Federated Retrieval, Agentic Media Environment and Software Defined Workflows) is an IBC (ibc) 2026 Accelerator project: an end-to-end intelligent media production framework that connects broadcaster archives, creative teams and AI agents. One Small Crawl (one-small-crawl) is a proof-of-concept / showcase video produced inside this project.

Overview

FRAMES aims to prove an end-to-end intelligent media production framework that dissolves traditional silos by standardising archival asset management and their metadata and enforcing Zero Trust security. Captured and archive material is transformed in a structured, interoperable environment where Agentic AI can automate tagging, enhancement, and semantic retrieval. Ontology for Media Creation-driven agents work alongside creators who stay in-the-loop to ensure authenticity while remixing archive and newly captured content (omc-ontology, ai-agent-pipeline).

Core challenge objective

Use a knowledge graph and a common semantic model to increase visibility into the production process, allowing clearer and better tradeoffs between creativity, time, and budget.

Approach

  • Starts from a script produced by the team; MovieLabs OMC is used to breakdown scripts and expose narrative and production information to downstream applications.
  • Archive content is tagged per a pre-defined filmmaking dictionary (camera shots and angles, look and feel, narrative).
  • A swarm of locally-running agents performs post-production operations (upscaling, in-painting/out-painting, generative extension, noise removal, audio cleanup) to enhance retrieved content.
  • Event-driven collaboration: a rejected capture makes agents notify sound, colour and VFX specialists with precise improvement instructions (the POC includes a simulated late script-break change).
  • Hybrid capture-and-archive system; the POC delivers a short video as the final project deliverable.

Consortium

  • Champions: EBU (ebu), ITV, ETC/USC, XRBB, BBC, RAI (rai), MovieLabs (movielabs), INSA National Institute of Applied Science (insa, INSA Hauts-de-France Valenciennes), The Global Creative and Security Community.
  • Participants: AMD (amd)/HP (hp), Enigmatic, Streamline Media (streamline-media), eddie-ai, AIMICI. Supporting participant per the IBC2026 press release: Holli (holli).
  • Proposed by RAI, EBU and MovieLabs; kickstart day pitch at IBC2026.

Relationship to One Small Crawl

  • One Small Crawl (production id IBC2026-SNAIL-001) is the animated short made as the FRAMES showcase/POC deliverable. Its archive-plus-AI hybrid pipeline and agent-orchestrated workflow mirror the FRAMES architecture: ARC retrieval agent AGT-016 (federated retrieval from RAI archive), asset_origin tagging, era-locked look applied in post, and the OMC export ingested into this wiki (per-shot OMC JSON, agent registry, OTIO). See one-small-crawl, archive-footage-reuse, animatic-workflow.

References (web)

  • show.ibc.org - “Accelerator Project 2026: FRAMES: Federated Retrieval, Agentic Media Environment and Software (Defined Workflows)” and the IBC2026 Accelerator press release

FRAMES - IBC2026 Accelerator project (no Wikipedia article; IBC official page)