Iyuno introduces multi-agent AI for media localization

Iyuno, a major media localization company, has introduced a multi-agent engineering approach for its CLOE system to address narrative continuity in long-form video. Rather than relying on massive, monolithic AI models, the company utilizes a network of hyper-specialized micro-agents designed for specific tasks such as character relationship mapping, emotional intent, and brand compliance.

The system operates on three core principles: Vertical Multi-Agent Orchestration, High-Density Low-Token Prompting, and Persistent Graph Memory. By synthesizing raw video, audio, and scripts into a structured knowledge graph, the agents work from compressed context vectors. This method aims to reduce token consumption and inference costs while ensuring that understanding compounds across entire titles, seasons, or franchises.