5 Key Ways Custom Catalogs and Profiles Are Revolutionizing Enterprise MCP Adoption

Introduction

The Model Context Protocol (MCP) is rapidly becoming the backbone of enterprise AI tooling, but managing and distributing MCP servers at scale has always been a challenge. Today, two powerful new capabilities—Custom Catalogs and Profiles—are changing the game. These tools allow organizations to curate approved collections of MCP servers, while giving developers the flexibility to build, share, and reuse their own tool configurations across projects. In this article, we break down the five most important things you need to know about these innovations, including how to create your own custom catalog, define portable profiles, and unlock new levels of productivity and governance in your AI workflows.

1. What Are Custom Catalogs and Why Do They Matter?

Custom Catalogs are curated, organization-wide collections of MCP servers. Instead of each team member hunting for MCP servers across the open internet, you can publish a single, approved catalog that includes internally built servers, community sources, and servers from Docker's MCP Catalog. This centralization brings three key benefits: security (only vetted servers are available), discoverability (teams can find the right tools faster), and trust (all servers meet your organization's standards). Think of it as an app store for AI tooling—but one you control entirely. With Custom Catalogs, enterprises can finally manage MCP deployment at scale, reducing both the overhead of manual onboarding and the risk of unauthorized tool usage.

2. MCP Profiles: Portable Toolkits for Every Developer

While Custom Catalogs focus on organization-wide curation, MCP Profiles are all about individual and team-level flexibility. A Profile is a named, portable grouping of MCP servers that you can define once and reuse across multiple projects, environments, or even share with colleagues. For example, you might create a "Data Analysis" profile that includes servers for querying databases, running statistical models, and generating visualizations. Profiles solve a common pain point: developers often spend hours configuring MCP servers for each new project. With Profiles, you simply reference a named configuration, and all the servers are automatically wired up. They also make collaboration seamless—your team can clone your profile and get the same setup in seconds. Profiles are designed to be extensible, so expect even more powerful features in the future.

3. Step-by-Step: Creating Your First Custom Catalog

Building a Custom Catalog is straightforward. Start by identifying the MCP servers your team needs—both from trusted public sources and your own internal tools. For each server, create a metadata file (usually in YAML format) that includes its name, title, type (server or client), and the location of its Docker image (e.g., your-dockerhub-id/mcp-server@latest). Once you have all the metadata files, combine them into a single catalog manifest. This manifest acts as the index for your catalog. You can then host this manifest in a private repository or a shared file system. Tools like Docker Desktop and the CLI can import the catalog by pointing to the manifest URL. For security, you can sign the catalog to prevent tampering. The result: a trusted, company-wide list of MCP servers that every developer can access with a single command.

4. Combining External and Internal Servers in One Catalog

The real power of Custom Catalogs comes from mixing servers from different sources. Suppose your team uses the official Docker MCP Catalog for standard tools like web scraping and PDF processing, but also has a custom internal server for handling proprietary APIs. With a Custom Catalog, you can include both in the same manifest. For example, you might pull the web-scraper server from the Docker Catalog and add your own roll-dice server (a simple MCP server that communicates over stdio and is built as a Docker image). By listing all these servers in one catalog, developers no longer need to search multiple repositories or remember complex URLs. They simply install the catalog and all approved servers appear in their IDE or chat interface. This hybrid approach balances innovation (internal tools) with stability (curated third-party services).

5. How Profiles Enable Faster Collaboration and Portability

Profiles take the concept of portability to the next level. Imagine you've built a profile called "Full-Stack AI Development" that includes servers for code generation, bug analysis, and documentation. You can share this profile with a colleague by exporting it as a file or referencing it from a shared registry. When they import it, all the associated MCP servers are automatically configured—no manual setup required. This is especially valuable in team settings where everyone needs to use the same tooling. Profiles also work across environments: you can define a profile for development, a different one for testing, and another for production, each with its own set of servers and configurations. The ability to version profiles (e.g., profile-v2.1) ensures that changes are trackable and reversible. In short, Profiles turn the chaotic process of MCP setup into a reproducible, shareable, and scalable practice.

Conclusion

Custom Catalogs and Profiles represent a major leap forward in enterprise MCP adoption. By providing centralized curation and portable configuration management, they address the two biggest barriers to scaling AI tooling: governance and usability. Organizations can now enforce security policies while giving developers the freedom to innovate. Whether you're a small startup or a large enterprise, adopting these tools will streamline your MCP workflow, reduce friction, and accelerate AI development. Start by creating a Custom Catalog for your team, then experiment with Profiles to see how they simplify your daily work. The future of MCP management is here—and it's more organized than ever.