TLDR: Teleport, a leader in infrastructure identity, has announced enhanced support for the Model Context Protocol (MCP), a critical step in securing interactions between Large Language Models (LLMs) and sensitive enterprise data. This integration extends Teleport’s robust identity, access control, and audit capabilities to AI workflows, addressing new security challenges posed by the rapid adoption of LLMs in production environments.
SAN FRANCISCO, CA – Teleport, the Infrastructure Identity company, has significantly advanced enterprise AI security by announcing comprehensive support for the Model Context Protocol (MCP). This strategic move aims to safeguard the increasingly critical interactions between Large Language Models (LLMs) and an organization’s underlying infrastructure data.
The rapid integration of LLMs into core enterprise workflows has introduced a new frontier in access and data governance. While MCP, pioneered by Anthropic and adopted by industry giants like Microsoft, OpenAI, and Cloudflare, facilitates structured interactions between LLMs and various data sources, it simultaneously presents novel security challenges that traditional models are ill-equipped to handle. The need for controlled and meticulously audited data access for AI systems has become paramount.
Teleport’s solution extends its established Infrastructure Identity Platform to MCP workflows, unifying access control across human, machine, and AI identities. This approach eliminates the reliance on static credentials, enforces task-based authorization, and ensures audit-ready visibility for all LLM-powered systems. The company’s new MCP support guarantees that LLM interactions with infrastructure data strictly adhere to Teleport’s rigorous identity, access control, governance, and audit standards already applied to other infrastructure technologies.
Key features of Teleport’s MCP support include: Strict Access Control: LLMs are granted access only to explicitly authorized resources. Principle-of-Least-Privilege: Authorization is tightly scoped, ensuring LLMs perform only actions precisely permitted by user roles. Comprehensive Audit Trails: Every LLM data access attempt, whether successful or denied, is meticulously logged, providing complete transparency and accountability.
Ev Kontsevoy, CEO and co-founder of Teleport, emphasized the importance of this development, stating, “Adopting new technology is always hard, because security is always a bottleneck. It is rare that a technology as impactful as AI comes along, and so the pressure to deliver innovation at speed is significant. By seamlessly enforcing strict access controls and comprehensive auditing through Teleport, organizations can confidently adopt LLMs, unlocking innovation while adhering to existing security and compliance frameworks.”
Stephanie Walter, Analyst in Residence at HyperFRAME Research, echoed this sentiment, noting, “As enterprises embed LLMs deeper into production environments, securing the interactions between models and infrastructure data becomes mission-critical. Teleport’s approach to securing MCP is a smart and timely response to a rapidly evolving and complex security challenge.”
Also Read:
- New AI Vulnerability ‘IdentityMesh’ Exposes Cross-System Exploitation Risks
- Tenable Unveils AI Exposure to Fortify Enterprise Generative AI Security
This MCP support underscores Teleport’s commitment to accelerating engineering velocity while simultaneously hardening infrastructure resiliency, facilitating rapid technological advancement, and safeguarding enterprise data. The early access release of Teleport’s MCP support was made available in June, with further developments including the Teleport Infrastructure Identity Platform’s availability in the new AWS Marketplace AI Agents and Tools category as of July 16, 2025, and the launch of Identity Security Capabilities on July 15, 2025, to eliminate hidden infrastructure risk.
