Linux Foundation Launches agentgateway: Unified Open-Source Gateway for AI Agents and Services
Key Takeaways
- ▸Single unified gateway handles service traffic, LLM routing, MCP servers, and agent-to-agent communication, eliminating the need for multiple specialized gateways
- ▸Supports all major LLM providers (OpenAI, Anthropic, Gemini, Bedrock) plus self-hosted models with per-team token budgets, semantic caching, and prompt redaction
- ▸Native support for Model Context Protocol and agent-to-agent communication enables tool discovery, versioning, RBAC, and audit trails without additional wrappers
Summary
The Linux Foundation has announced agentgateway, an open-source HTTP and gRPC gateway designed to unify traffic management across traditional services, AI language models, and Model Context Protocol (MCP) tools in a single data plane. This addresses a critical gap in AI infrastructure where organizations previously needed to stitch together separate gateways for different protocol types. The project supports routing to major LLM providers including OpenAI, Anthropic Claude, Google Gemini, and self-hosted models, while natively supporting MCP servers and agent-to-agent communication.
The gateway emphasizes security and observability by default, incorporating OpenTelemetry tracing, mTLS with automatic rotation, OIDC authentication, and OPA-based policy evaluation across all traffic types. It includes smart routing for self-hosted GPU pools that optimizes for latency, cost, and model capabilities. The project is backed by a broad coalition of industry leaders—Microsoft, Apple, AWS, Alibaba, Adobe, Cisco, Salesforce, Huawei, and others—positioning it as a collaborative effort to build neutral, interoperable infrastructure for the emerging agent ecosystem.
- Built-in security (mTLS, OIDC, OPA policies) and observability (OpenTelemetry, per-call tracing) with tamper-evident audit logs
- Broad industry backing from Microsoft, AWS, Apple, and others signals consensus on the need for open, vendor-neutral AI agent infrastructure
Editorial Opinion
agentgateway addresses a genuine infrastructure gap at a critical inflection point for AI. As enterprises scale from experimental chatbots to sophisticated multi-agent systems, they need unified governance, visibility, and control—not a patchwork of specialized gateways. The Linux Foundation's stewardship and the breadth of contributing companies suggest this could become foundational open infrastructure for the agent economy. This is exactly the kind of collaborative, vendor-neutral project needed to accelerate responsible AI adoption at enterprise scale.
