IBM Launches MCP Gateway to Merge and Manage AI Tools

IBM has announced a new MCP Gateway to simplify how developers build and scale AI-driven systems. It was announced by Armand Ruiz, VP of AI Platforms at IBM, in a LinkedIn post earlier this week.

Built on FastAPI, the MCP Gateway is designed to act as a unified entry point for the Model Context Protocol (MCP), specifically adhering to the official MCP 2025-03-26 protocol. Acting as a true gateway, it centralizes tool, resource, and prompt registries. 

It allows developers to bring multiple MCP servers under one endpoint, sitting in front of existing MCP Servers or REST APIs to wrap them into MCP-compliant tools and support various transport protocols, including HTTP/JSON-RPC, WebSocket, Server-Sent Events (SSE), and standard input/output.

One of its standout features is the ability to manage tools, prompts, and resources entirely using JSON-Schema validation. It also comes with an admin UI that includes built-in authentication, observability, and transport layer management.

According to Ruiz, this launch is particularly relevant for teams working on agent-based systems or orchestrating multiple AI tools within enterprise environments. He credited IBM Consulting for helping shape the tool based on real-world enterprise use cases.

However, it’s important to keep in mind that this is not the first AI update from IBM. Recently, the company partnered with Inclusive Brains to integrate AI, quantum computing, and neurotechnologies.

Apart from that, the company has announced the opening of Watsonx AI Labs in New York City. This new facility is positioned as an accelerator for AI builders, startups, and enterprises, providing access to advanced AI tools, resources, and expertise.