v0.8 : Technical Preview
This is the first public release of the Nutanix V4 API MCP Server. This release is a Technical Preview and is not designed or supported for production workloads.
What this is
Nutanix V4 API MCP Server is an open-source bridge between AI assistants and Nutanix Prism Central, exposing your infrastructure to natural language with a V4 API surface that automatically reflects the version your Prism Central cluster is running.
AI assistants like Claude and Cursor are useful for writing code and answering questions. Until now, they had no direct path to your infrastructure. Every Nutanix operation still required navigating Prism Central's UI, writing scripts, or calling the V4 REST API directly. This server changes that by giving AI assistants the ability to discover, understand, and execute Nutanix V4 API operations against your actual cluster, with your credentials, with every action logged.
The shift this enables is concrete: infrastructure operations move from context-switching between documentation and tooling, to a conversation. New engineers can explore what's possible without knowing the API surface upfront. Experienced engineers automate faster by describing intent rather than implementation. Security teams get a structured audit trail of every AI-initiated action.
This is a foundation. As AI assistants mature and this server moves toward general availability, the range of what is expressible in a prompt grows with it.
What's in this release
V3 to V4 Migration Assist
Paste a script built on V3 REST APIs. With access to the full V4 operation index, correct parameter contracts, and working code samples, an AI assistant can identify equivalent V4 operations and produce accurate rewrites, turning days of API archaeology into a single conversation.
Adaptive, Deployment-Aware API Surface
Unlike static API wrappers pinned to a generic latest version, this server resolves the V4 API surface available on your Prism Central deployment at connection time. Teams running different Prism Central versions each operate against schemas that match what their cluster actually supports, with no version mismatch and no silent incompatibilities.
Infrastructure Intent, Executed
Describe what you need: provision a VM, query protection policy state, sequence a multi-step remediation. The server resolves the correct V4 operation(s), required and optional fields, and RBAC permissions before any call is made. End-to-end infrastructure workflows execute entirely through natural language, with no intermediate tooling or manual API construction.
Governed Access, Out of the Box
Read-only enforcement is on by default. Write operations require an explicit opt-in, making the server safe to connect to sensitive or production-adjacent environments from day one. Every operation is audit-logged with method, path, and status, giving security and compliance teams a machine-readable record of every AI-initiated action. Air-gapped and restricted deployments are supported; the server requires no internet access at runtime.
Precision Search Across 1,000+ V4 API Operations
The server indexes the full V4 API surface at startup, covering 1,000+ operations across compute, networking, storage, data protection, security, and more. Queries return operation identifiers, parameter contracts, and permission requirements scoped to what is relevant and what your cluster supports, with the accuracy and efficiency that AI workflows at scale require.
Getting started
See the quickstart guide to connect your AI assistant to Prism Central in minutes.
Feedback
File a bug, feature request, or documentation issue using the GitHub Issue templates. This is a preview and your feedback shapes what ships next.