Cloudinary Launches Agent Experience Layer to Automate Visual Media Workflows

Cloudinary has launched an Agent Experience layer and a suite of specialized AI agents to enable agent-native development of visual media workflows. This infrastructure allows AI agents to autonomously provision environments and execute complex media tasks using natural-language prompts.

At the center of the release is the Agent Experience layer, which introduces "Claimable Clouds." These allow AI agents to self-provision temporary environments to integrate media functions without requiring human account setup or manual configuration. To support these operations, Cloudinary introduced five specialized agents: a Taxonomy Agent for classification, a Search Agent for natural-language discovery, a Workflow Agent for automation, a Moderation Agent for brand and compliance review, and a Coordinator Agent for orchestration. The platform also supports the Model Context Protocol (MCP), providing AI-optimized SDKs and tooling to ensure LLMs produce predictable outputs when interacting with Cloudinary APIs.

The Shift to Agentic Infrastructure

This move comes as enterprises transition toward agentic software development, where AI agents—rather than just human developers—discover services and provision infrastructure. Traditionally, SaaS onboarding has been a bottleneck for autonomous agents because it requires human intervention for billing and account creation. By treating visual media as structured assets with machine-readable metadata, Cloudinary is removing these manual hurdles.

Tal Lev-Ami, Co-Founder and CTO at Cloudinary, noted that software is no longer built exclusively by humans writing code, stating that developers now work alongside AI agents that evaluate architectures and build integrations. This shift is particularly urgent for media teams; Rob Daynes, General Manager of Assets at Cloudinary, observed that these teams face increasing pressure from a growing volume of content and channels while operating with limited resources.

Industry Implications

This launch marks a transition from AI-assisted tools to agent-native infrastructure. By eliminating the human-in-the-loop requirement for initial provisioning and providing a governed framework, companies can now automate high-volume, multi-step visual workflows—such as global product launches—that previously required extensive custom code and manual oversight. This capability is being rolled out to a customer base of over 11,000 clients, which includes major brands such as Adidas, Etsy, Fiverr, and Mattel.

What to Watch

As the industry adopts the Model Context Protocol, the interoperability between Cloudinary's agents and broader enterprise software stacks will be a key metric for success. Future developments will likely focus on how these autonomous agents handle increasingly complex brand governance across diverse global markets without human intervention.