Agentgateway Becomes Fourth Hosted Project Under Agentic AI Foundation

Summary

Agentgateway, an open-source gateway infrastructure designed specifically for agentic AI systems, has joined the Agentic AI Foundation (AAIF) as its fourth hosted project under the Linux Foundation. The project addresses a critical gap in infrastructure designed for modern AI workloads, where agents call tools, models route across providers, and workflows span multiple APIs, MCP servers, and databases. Unlike traditional web infrastructure, agentgateway provides a unified control plane and proxy data plane for managing MCP traffic, Agent-to-Agent communication, LLM inference, REST APIs, and gRPC services.

Agentgateway brings a comprehensive set of capabilities to help organizations operate agentic systems safely and reliably at scale. It provides unified security controls including JWT authentication, API key auth, RBAC, external authorization, and mTLS; built-in observability with metrics, tracing, and AI-specific logging; and governance features such as rate limiting, content-based routing, and model aliasing. The project supports MCP virtualization and enables platform-agnostic deployment across bare metal, VMs, containers, and Kubernetes.

The move to AAIF signals growing industry recognition that agentic AI requires different operational and infrastructure approaches than traditional AI workloads. By joining alongside standards like MCP and other open-source tooling, agentgateway becomes part of a broader ecosystem focused on open, neutral governance of critical agentic AI infrastructure under the Linux Foundation.

• Available open source under Apache 2.0 with support for MCP, Agent-to-Agent communication, and platform-agnostic deployment