Buoyant is extending Linkerd to support the Model Context Protocol, enabling emerging agentic AI workloads on a unified, zero-trust, open source service mesh.
Buoyant, the creator of Linkerd, the open source service mesh for cloud-native and Kubernetes environments, has announced upcoming support for the Model Context Protocol (MCP). This move positions Linkerd as the first open source service mesh to natively support MCP, offering a unified and secure foundation for both traditional microservices traffic and agentic AI-driven communication.
Developed by Anthropic, MCP is an open standard that allows AI models to connect with external data sources and tools through secure, two-way channels. Unlike conventional API-based communication, agentic AI exchanges are persistent, stateful and often multi-step in nature. These workloads can introduce novel security risks, generate volatile traffic spikes and increase operational complexity. Many enterprises lack the infrastructure maturity required to manage these forms of traffic safely.
By adding MCP support, Linkerd will extend its lightweight security, observability and traffic management capabilities to agentic AI workflows. This includes metrics on resource and tool usage, latency and failure rates, along with fine-grained, identity-based access controls built on Linkerd’s existing zero-trust framework. The aim is to provide a single mesh layer that maintains consistent policy enforcement across all workloads.
“Enterprises are eager to innovate with AI, but they cannot do so at the expense of their security posture and application reliability,” said William Morgan, CEO of Buoyant. “Linkerd solves this problem by extending its proven capabilities to MCP traffic. We are giving organisations the tools to accelerate their usage with confidence.”
Full MCP support will be available in 2026 across both the open source Linkerd distribution and Buoyant’s enterprise offering.
