ExpressVPN just gave AI agents the power to control your connection

• ExpressVPN launches industry's first Model Context Protocol (MCP) server
• The beta allows AI agents to interact directly with its VPN desktop apps
• The feature is opt-in, runs locally, and operates under a strict no-log policy

ExpressVPN has announced the beta launch of its Model Context Protocol (MCP) server, becoming the first major provider to allow AI tools to interact directly with its desktop applications.

Announced today (March 5), this addition bridges the gap between the best VPN services and the booming world of AI agents. By adopting the open standard introduced by Anthropic in late 2024, ExpressVPN is effectively giving developers the ability to control their network security using natural language or automated scripts within their coding environments.

Previously, if a developer using an AI assistant needed to test a geo-specific API or verify a secure connection, they had to leave their workflow to manually toggle settings in the VPN app. Now, compatible AI tools like Anthropic's Claude Code and OpenAI's Codex can read the VPN's status, switch regions, and troubleshoot connectivity issues on behalf of the user.

Bringing AI to network security

The Model Context Protocol (MCP) is an open standard designed to connect AI systems with external data and tools. While many developer platforms have adopted it since its introduction, network infrastructure has largely remained a manual affair — until now.

ExpressVPN’s implementation creates a "strictly local bridge" between the AI tool and the ExpressVPN desktop app on macOS, Windows, or Linux. This means developers can issue commands in plain English, such as "switch to a US server" or "check if my connection is secure", and the AI agent executes them via the MCP server.

Shay Peretz, COO of ExpressVPN, says the move is about removing friction for technical users who live in their terminals.

"Developers are increasingly relying on AI agents to automate complex tasks, yet their network environment has remained manual," Peretz said. "Our MCP server brings VPN control into those workflows, enabling secure testing, troubleshooting, and region switching without interrupting their workflow."

Peretz added that the company sees MCP becoming a "foundational standard for how AI systems interact with external tools."

Private by design

Allowing an AI to control network settings might raise eyebrows regarding security, but ExpressVPN emphasizes that the feature is "private and secure by design."

Crucially, the integration is disabled by default; users must actively opt in via the desktop app settings. Furthermore, the MCP server operates on a fixed command allowlist. It cannot access account credentials or session information, ensuring that the AI agent doesn't get free rein over sensitive user data.

The company also confirmed that the new feature operates under its strict no-logs policy. "Activity logs, connection logs, and destination data are not collected or stored," said ExpressVPN.

Who is this for?

While the average user might not need their AI chatbot to swap VPN servers just yet, this is a significant quality-of-life upgrade for software engineers and QA testers.

The beta version enables developers to automate tasks such as verifying secure connections before running scripts, testing geo-sensitive APIs across multiple regions, and diagnosing network issues through AI-assisted workflows. It essentially removes the need to context-switch between a code editor and the VPN client.

The MCP server beta is available starting today, March 5, 2026, for individual users and ExpressVPN for Teams customers on all major desktop platforms. You can find more details on the ExpressVPN features page or their support guide.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!