X Launches Hosted MCP Servers to Connect Cursor, Claude, and Other AI Tools

X (formerly Twitter) has officially launched hosted Model Context Protocol (MCP) servers, enabling AI development tools such as Grok Build, Cursor, and Claude Desktop to seamlessly connect with the platform’s API and documentation. Announced on Tuesday, the move positions X as an early adopter of Anthropic’s open MCP standard for agentic AI integrations at scale.

Model Context Protocol (MCP) is an open standard developed by Anthropic that allows AI tools and agents to interface with external services in a structured, permission-controlled manner.

By hosting MCP servers natively, X eliminates much of the configuration overhead that previously made API integrations complex for developers. Instead of manually managing authentication flows and endpoint configurations, developers can now connect their AI tools to X’s data layer with minimal setup.

This is particularly significant for security researchers and threat intelligence analysts who rely on real-time social data X’s API has long been a primary source for tracking emerging threats, threat actor communications, and vulnerability disclosures.

How Developers Connect

Connecting to X’s hosted MCP servers is straightforward. Developers must:

• Enable OAuth authentication through the X Developer Portal
• Install and configure the open-source xurl CLI tool
• Select the desired API features to expose to their AI agent or tool

Once connected, tools like Cursor and Claude Desktop gain programmatic access to powerful X platform features including full-archive search, trending topics, and bookmarking.

The integration supports agentic workflows, meaning AI tools can autonomously query, retrieve, and process X data as part of multi-step reasoning chains without requiring manual developer intervention at each step.

X’s implementation respects existing rate limits and pricing tiers, which serves as a built-in throttling mechanism against abuse. OAuth-based authentication ensures that access is scoped and auditable, a critical requirement when AI agents are operating autonomously on behalf of users.

However, early adopters have flagged one notable gap: write operation auditing. While read access to X data through MCP appears well-structured, some developers note that better logging and auditing controls are needed for write-enabled workflows.

This is an important consideration from a security standpoint agentic systems with write access to social platforms introduce risks around unauthorized posting, data manipulation, or scope creep beyond intended permissions.

The launch was spearheaded by X’s developer team, with engineers Vardhan Agnihotri and Santiago Medina among those celebrating the release. Community feedback has been largely positive, with developers praising the potential for live data integration in AI-powered applications. The ability to ground AI responses in real-time X data is seen as a major step forward for building context-aware, up-to-date agents.

Full documentation for X’s hosted MCP servers is available at docs.x.com/tools/mcp.

 Strengthen Your SOC by Accelerating Threat Detection & Rapid Investigations. -> Integrate ANY.RUN With Your SOC Now.