Overview
The AgentRPC MCP Server is an optional component of the TypeScript SDK that implements the Model Context Protocol (MCP). It exposes an MCP-compliant interface that enables external AI models to interact with your registered tools and APIs through the AgentRPC platform. Start it with a secret, for example: ANGENTRPC_API_SECRET=YOUR_API_SECRET npx agentrpc mcp, which authenticates requests and launches MCP endpoints for discovery, invocation, and context-aware reasoning. The MCP server complements the broader AgentRPC workflow, where functions are registered and health-monitored by the central service (api.agentrpc.com), and clients receive OpenAPI SDK compatible tool definitions alongside an hosted MCP server for connecting to compatible agents. You can configure integration with third-party clients like Claude Desktop and Cursor by adding MCP server definitions to their config files. Designed to work in private networks and across languages, the MCP server enables cross-language access to tools and pairs with MCP/OpenAI SDK-compatible agents to form a flexible, interoperable AI tooling layer.
Features
Multi-language Support
Connect to tools across languages (TypeScript, Go, Python and .NET coming soon).
Private Network Support
Register functions in private VPCs with no open ports required.
Long-running Functions
Long polling SDKs allow function calls beyond HTTP timeout limits.
Full Observability
Comprehensive tracing, metrics, and events for complete visibility.
Automatic Failover
Intelligent health tracking with automatic failover and retries.
Framework Compatibility
Out-of-the-box support for MCP and OpenAI SDK compatible agents.
Who Is This For?
- Developers:Expose registered tools to external AI models via MCP using AgentRPC TypeScript SDK.
- AI teams:Integrate cross-language MCP-enabled tools into AI agents and automated workflows.
