AI in exchange: The Docker MCP Catalog collects containerized AI resources
With Docker MCP Catalog, the platform is launching a service for exchanging AI resources: simple download of servers and clients including authentication.
(Image: erstellt mit Dall-E durch iX.)
The container management platform Docker is announcing a new online directory for MCP tools for May: the Docker MCP Catalog. It collects AI servers and clients that can be connected via the Model Context Protocol (MCP). It is to be handled in the same way as the Docker Hub, which also serves as the basis for the infrastructure. The MCP tools are located in containers and can be integrated with just a few clicks or, conversely, offered as servers. Verification, authentication, and security are automated via the platform.
The Docker MCP Catalog is intended to be a central, simple and trustworthy point of contact for developers to find or share AI tools. For the launch, Docker is announcing over a hundred verified MCP servers on its blog from companies such as Stripe, Elastic, Heroku, Pulumi, Grafana Labs, Kong, Neo4j, New Relic and Continue.
(Image: Screenshot)
Similar to Docker Hub, the containerized tools can be integrated into your application with just a few clicks to load special AI functionalities. Clients for linking these servers are available at the start for Docker AI Agent, Claude, Cursor, VS Code, Windsurf, Continue and Goose. Here, too, installation should take just a few clicks, as a promotional video from Docker suggests.
The Catalog is based on the same infrastructure as Docker Hub, and companies can regulate access to resources for employees.
Videos by heise
Offer servers yourself
Docker provides the MCP Toolkit for developers who want to offer AI services themselves via MCP. Users use this to publish their models for all environments in which Docker is running. A special gateway is used to connect to the clients and an mcp terminal is used for build, run, and management. Commands for this are, for example
docker mcp server builddocker mcp catalogaddddocker mcp auth connect
Security in the system
In addition to the simple exchange of context from AI models, Docker emphasizes the security standards: providers can have themselves verified and receive a blue tick. Simple authentication and secrets management for users is carried out via the Docker account with OAuth or your token. The tools themselves work in isolation in the container.
(Image: Screenshot Video)
(who)
