Beyond Models: JFrog AI Catalog Evolves to Detect Shadow AI and Govern MCPs

When we first introduced the JFrog AI Catalog, it was our mission to provide the industry with a single system of record for governing the complex landscape of internal, open-source, and external commercial AI models. This foundational step was critical for enterprises to move from uncontrolled innovation to delivering AI with trust and confidence.

However, the AI landscape is ever-evolving. The challenge for today’s enterprise is already evolving beyond simply managing a library of known models. This rapid evolution is introducing new and urgent challenges, particularly the ungoverned usage of AI models and the rise of powerful automation agents.

This is why we are excited to announce the next evolution of the JFrog AI Catalog. We are expanding our platform to provide centralized, enterprise-grade governance over your entire AI ecosystem, because, ultimately, you can’t govern what you can’t see, and you can’t control what you don’t have a framework for.

The Rise of Shadow AI and MCP Servers

A Governance Nightmare: What is Shadow AI?

Developers and AI/ML teams are rightly focused on innovation, which often means quickly pulling in every tool available. This includes making direct API calls to external services like OpenAI, Gemini, and Anthropic, or pulling models from public repositories. Unfortunately, this is often done without any organizational oversight.

This uncontrolled usage creates a massive governance nightmare known as Shadow AI. Shadow AI is defined as the uncontrolled use of unmanaged model packages and external APIs, which creates significant security blind spots and compliance risks.

A New Risk Vector: Model Context Protocol (MCP) Servers

At the same time, a powerful new class of AI automation is emerging. These are agentic frameworks that often use Model Context Protocol (MCP) servers. While these technologies are poised to revolutionize business and development, they also represent a potent new risk vector if they are left unmanaged.

The Next Evolution of AI Catalog: Governance for the Full AI Ecosystem

We are thrilled to introduce the next evolution of the JFrog AI Catalog – delivering the centralized, enterprise-grade governance platform you need to address Shadow AI and emerging MCP server risks, and to finally control what was previously unseen and unmanaged.

Illuminate the Blind Spots With Shadow AI Detection

To combat the unchecked spread of Shadow AI, the JFrog AI Catalog now actively detects and inventories every AI model and external API call in use across your organization to give you a complete, centralized view of all models and models’ API calls.

How we eliminate the blind spots:

• Automatic Discovery: We instantly identify and catalog all unmanaged external AI API calls (like OpenAI, Gemini, and Anthropic) and models pulled from public repositories.
• Asset Vetting: Once an unknown asset is identified, the AI Catalog immediately brings it under governance, transforming an invisible risk into a vetted, managed asset.
• Policy Enforcement: You can enforce existing security and compliance policies, allowing you to automatically block non-compliant or malicious models and external APIs at the gate.
• Centralized Access with the AI Gateway: We provide a single, standardized connection for your developers, routing all AI consumption through the AI Gateway to bring your entire AI footprint under unified control.

Extend Proven Governance to MCP Servers

In the next few months, the JFrog AI Catalog will also extend its proven governance framework to the next wave of AI automation: MCP servers. We will be delivering a unified Model and MCP Registry, a single source of truth that allows you to manage and secure these new agentic assets with the same trusted controls you already use for models.

This new registry will be the critical first step toward secure and widespread adoption of MCP servers. From there, the AI Catalog will allow you to immediately bring these servers under governance, effectively turning a new and unknown risk into a vetted and managed asset. Once these agents are identified, you can:

• Apply your existing security policies to them.
• Automatically block non-compliant or malicious MCPs at the gate.
• Provide a secure, standardized connection for these agents through the AI Gateway.

This capability will effectively transform a new, high-risk vector into a managed, controlled, and trusted asset, allowing you to innovate safely with the most advanced AI frameworks.

The Trusted Source for Your Entire AI Supply Chain

When we launched the AI Catalog, we promised it would be The Trusted Source for Enterprise AI. Today, we are extending that promise to cover the complete AI lifecycle.

The JFrog Platform, with the newly expanded AI Catalog, is the only solution that provides a single system of record to secure and govern your entire supply chain: from the first line of code, to the external APIs you consume, to the most advanced AI agent.

To learn more about these new capabilities and see them in action, explore the AI Catalog product page or book a demo.