# Nutanix adding AI agent access bridge to its Cloud Platform > Source: > Published: 2026-08-10 17:02:58+00:00 # Nutanix adding AI agent access bridge to its Cloud Platform [Nutanix](https://www.blocksandfiles.com/ai-ml/2026/07/02/nutanix-providing-ai-agent-identity-access-management/5265784) is adding an open-source MCP server to its Cloud Platform software so that AI Agents can access it and use its resources. The Nutanix Cloud Platform (NCP) is KVM hypervisor-based, hyperconverged infrastructure software that is available both on-premises and in both public and private clouds, and supports [external storage](https://www.blocksandfiles.com/file/2026/07/28/nutanix-gets-dell-powerstore-support/5279629) as well as virtualising local storage on its servers. The [MCP](https://www.anthropic.com/news/model-context-protocol ) (Model Control Protocol) standard was open-sourced by Anthropic in late 2024 and provided a universal way for AI assistants and agents to connect to systems containing data they need to access. Thomas Cornely, EVP Product Management at Nutanix, said: “The MCP server is an exciting next step on our journey, allowing us to bring the power of autonomous AI agents to cloud management. By creating a secure gateway between AI tools and our platform, we are giving customers the confidence to safely use AI to operate and govern their hybrid multicloud environments.” Adding MCP access to a product widens its vulnerability surface. The MCP server functions as a bridge that can run commands, query databases, call APIs, read/write files, and return data on behalf of an AI agent. Nutanix says its MCP server acts as a secure passthrough between AI assistants, such as GitHub Copilot, Claude Code, and Cursor to safely translate plain-English requests into precise infrastructure API actions for an organization's hybrid multicloud environment. It “translates everyday language requests into precise, safe system actions while automatically enforcing existing corporate security and access policies. By creating a dedicated safety passthrough around AI tools, the solution allows IT teams to confidently automate routine operations and build custom AI-driven workflows.” Nutanix admits the vulnerability, saying enterprise AI tools are only as secure as the platform APIs powering them. Its MCP server builds directly on the Nutanix Prism V4 API Gateway, so that connected AI tools inherit the native governance and fine-grained access controls of NCP. It says IT operators can safely use AI assistants to analyze system health, surface diagnostics, and accurately stage automation workflows while maintaining operator oversight through 'human-in-the-loop' capabilities. The Nutanix Prism V4 API Gateway strictly handles all execution, governance, and security controls, and supports; Fine-Grained Role-Based Access Control (RBAC): Confines autonomous agents to executing calls only on specific APIs for which they have been granted explicit, granular authorization. Throttling and Metering: Implements built-in traffic regulation to help prevent agent swarm attacks while natively tracking API usage across automated workflows. Comprehensive Auditing: Generates detailed system logs to maintain visibility into which AI agent initiated which specific command. Asynchronous Task Management: Offloads long-running operations as asynchronous tasks, allowing autonomous agents to efficiently track progress and execution status until completion. If teams can configure precise security permissions through a single interface, or build their own custom AI agents to automate and govern their Nutanix infrastructure from end to end. Developers can give AI coding assistants exact system blueprints up front and generate platform-ready scripts in their preferred languages – including Python, Go, Java, JavaScript, PowerShell, curl, or any REST and JSON-compatible format. Nutanix has already launched AI agent [identity access management](https://www.blocksandfiles.com/ai-ml/2026/07/02/nutanix-providing-ai-agent-identity-access-management/5265784). Customers can download the MCP server for NCP from [here](https://developers.nutanix.com/) and use it with any of their own AI agents.