The Model Context Protocol has become established as a means of communication for LLMs with external systems. Solo.io is launching MCP Gateway within kgateway for this purpose. This new gateway makes the integration and management of AI agents with MCP-compatible toolchains considerably easier for the delivery of modern, cloud-native applications.
Solo.io emphasizes the widespread adoption of its tool. Companies such as Domino’s Pizza, ParkMobile and Vonage already use kgateway in production environments for API traffic management, stability and security. With the new MCP Gateway, organizations can now easily implement advanced AI gateway applications. Due to the growth of AI workloads and the increasing need for traffic management for LLM providers, gateways play an increasingly crucial role in controlling and securing traffic in all directions.
Standard protocol for AI tools
MCP was originally developed by Anthropic, but is also used by OpenAI and open-source AI players. It offers a standardized method to connect LLM’s with various data sources and tools. The protocol is quickly developing into the standard for interoperability between AI agents and tools. The layer on top of MCP, however, suffers from a proliferation of options. AI clients and agents are dealing with ‘tool sprawl’, in Solo.io’s words, where there are thousands of options, each of which wants to become the standard for connectivity between systems.
With MCP Gateway, developers and platform teams can discover, secure and federate multiple MCP tools and tool servers in a virtualized MCP server. This provides developers with a single, secure MCP tool registration and access point, regardless of the number of tools used by AI agents and applications.
Benefits of MCP Gateway
“While MCP has rapidly emerged as the de facto standard protocol for accessing tools from agents, building MCP integration capabilities into every application client and agent is error-prone and cumbersome for AI development teams,” said Keith Babo, Chief Product Officer, Solo.io. “MCP Gateway eliminates the need to manually connect these agents and tools by consolidating MCP servers into one place, saving teams countless hours and eliminating undifferentiated heavy lifting.”
Important use cases
MCP Gateway offers several important use cases, including:
Simplification of tool onboarding with automated detection and registration of MCP tool servers.
A centralized registry of MCP tools for developers, regardless of the location of the heterogeneous tool servers.
● Access to every MCP tool via a single endpoint with innovative MCP multiplexing that transforms an entire ecosystem of thousands of tools into a virtualized MCP tool server.
● Directly secure MCP tool server implementations for consistent authentication and authorization checks for multi-tenant use.
● In-depth insights and observability in AI agent and tool integrations with centralized metrics, logging and tracing for all tool calls.
MCP Gateway is part of kgateway and is completely open source as a Sandbox project under the Cloud Native Computing Foundation.