Novita MCP Server

The Novita MCP Server is a lightweight, cloud‑agnostic bridge that lets AI assistants such as Claude or Cursor talk directly to the Novita AI platform. By exposing a standardized Model Context Protocol interface, it removes the need for developers to write custom SDK calls or manage REST endpoints. Instead, a single MCP server handles authentication, request routing, and response formatting, allowing AI agents to issue natural‑language commands that are translated into concrete GPU instance operations.

At its core, the server provides a unified view of Novita’s GPU‑instance ecosystem. Developers can list clusters and products, inspect the state of individual instances, and perform lifecycle actions—create, start, stop, restart, or delete. It also supports template management for reusable instance configurations, container‑registry authentication, and network storage provisioning. These capabilities give AI assistants the same level of control that a seasoned DevOps engineer would expect, but through conversational prompts.

The value for AI‑centric workflows lies in the seamless integration with existing MCP clients. Once configured, a user can type “Create a new GPU instance” and the assistant will prompt for parameters, translate them into an API call, and return a ready‑to‑use instance ID. This eliminates context switching between the assistant’s chat interface and external dashboards, speeding up experimentation cycles for machine learning teams. The server’s beta focus on GPU management is already a powerful tool for rapid prototyping, hyper‑parameter sweeps, or temporary compute bursts that would otherwise require manual provisioning.

Real‑world scenarios include data scientists leveraging Claude to orchestrate GPU clusters for training large language models, or software engineers using Cursor to spin up isolated environments for continuous integration pipelines. Because the MCP server abstracts away authentication (via an API key) and handles all HTTP intricacies, developers can focus on higher‑level logic while the assistant manages infrastructure state. As Novita expands its resource catalog, the server will grow to support additional types—storage volumes, networking rules, or even inference endpoints—making it a scalable foundation for AI‑driven DevOps.