TechPrufer
Get started
OpenAI
OpenAI logoOpenAI · Feature trail

Secure MCP Tunnels

1 post · Jun 2026

What this trail covers

This feature thread tracks the evolution of OpenAI's capabilities for securely connecting private MCP servers to hosted AI products. Initially, the focus was on enabling private servers to be reachable without exposing them to the public internet. This involved developing the Secure MCP Tunnel, which allows customers to run a client within their private environment that establishes an outbound HTTPS connection to OpenAI. The client receives MCP requests, forwards them to local servers, and returns responses and notifications through the same connection. This post extends the capability to support REST APIs via Harpoon, allowing controlled access to non-MCP private services without opening inbound network access.

All OpenAI trails
  1. Making private MCP servers reachable without making them public | OpenAI DevelopersThis post introduces Harpoon, an extension of the Secure MCP Tunnel that enables controlled access to private REST APIs. Harpoon allows customers to register labeled targets on the tunnel client, enabling OpenAI-side callers to invoke these labels through Secure MCP Tunnel. The actual HTTP request originates from inside the customer environment. This extension maintains the narrow connectivity model, bounding calls by customer-owned target registration, allowed methods, response-size limits, timeouts, redirect behavior, and tunnel access controls, providing a controlled path to private APIs without inbound network access.