AgentGateway
ActiveDescription
Next generation agentic proxy for AI agents and MCP servers. Provides unified traffic management, routing, and security control.
Key Features
- Unified LLM gateway with OpenAI-compatible API routing to OpenAI, Anthropic, Gemini, Bedrock and more
- MCP gateway with tool federation, stdio/HTTP/SSE transports, OpenAPI integration and OAuth authentication
- A2A gateway enabling secure agent-to-agent communication with capability discovery and task collaboration
- Multi-layered guardrails with regex, OpenAI moderation, AWS Bedrock Guardrails, and custom webhooks
- Security features including JWT/API key/OAuth auth, RBAC with CEL policy engine, rate limiting and TLS
- Kubernetes-native inference routing based on GPU utilization, KV cache, LoRA adapters, and queue depth
Use Cases
💡 Centralized traffic management and security governance for multi-agent AI deployments
💡 Unified API gateway for routing requests across multiple LLM providers with failover
💡 Agent-to-tool connectivity via MCP protocol with authentication and observability
💡 Kubernetes-based AI inference workload routing and resource optimization
Categories
Quick Start
Visit agentgateway.dev/docs/quickstart for the standalone quickstart guide. For Kubernetes, follow the guide at agentgateway.dev/docs/kubernetes/latest. Install agentgateway via binary release or Docker. Configure your LLM providers, MCP tool servers, and security policies. The built-in UI helps explore agent-to-agent and agent-to-tool connections.