Skip to main content
Glama
novitalabs

Novita MCP Server

Official
by novitalabs

create-template

Create reusable templates for GPU instances or serverless deployments on the Novita AI platform by defining Docker images, ports, environment variables, and configuration settings.

Input Schema

TableJSON Schema
NameRequiredDescriptionDefault
templateYes

Implementation Reference

  • The inline async handler function for the 'create-template' tool. It calls novitaRequest to POST the parameters to the /template/create endpoint and formats the result as a text content response.
    }, async (params) => { const result = await novitaRequest("/template/create", "POST", params); return { content: [ { type: "text", text: JSON.stringify(result, null, 2), }, ], }; });
  • Zod input schema defining the structure of the 'template' object parameter, including fields like name, readme, type, channel, image, ports, env, etc., with descriptions and validations.
    template: z.object({ name: z .string() .min(2) .nonempty() .trim() .describe("Name for the template, must be unique."), readme: z .string() .max(20480) .optional() .describe("Template README content (in Markdown format)"), type: z .enum(["instance", "serverless"]) .default("instance") .describe("Type of template"), channel: z .enum(["private", "community"]) .default("private") .describe("Channel for the template"), image: z .string() .nonempty() .max(500) .trim() .describe("Docker image address for instance startup"), imageAuth: z .string() .trim() .optional() .describe("ID of the container registry auth. Required only when the channel is private, or else keep it empty. You can use the `list-container-registry-auths` tool to check the ID if you're not sure."), startCommand: z .string() .max(2047) .optional() .describe("Command to run when the instance starts"), rootfsSize: z .number() .min(10) .describe("Root filesystem size (/Container disk size) in GB"), ports: z .array(z.object({ type: z.enum(["http", "udp", "tcp"]).describe("Type of exposed port"), ports: z.array( z.number() .min(1) .max(65535) .describe("Exposed port numbers, maximum of 10") ).max(10).describe("Exposed port numbers, maximum of 10"), })) .describe("Ports to expose"), env: z .array(z.object({ key: z.string().nonempty().describe("Environment variable key").max(2048), value: z.string().describe("Environment variable value").max(2048), })) .optional() .describe("Environment variables"), }),
  • src/tools.ts:349-419 (registration)
    The server.tool() call within registerTemplateTools() that registers the 'create-template' tool, specifying its name, input schema, and handler function.
    server.tool("create-template", { template: z.object({ name: z .string() .min(2) .nonempty() .trim() .describe("Name for the template, must be unique."), readme: z .string() .max(20480) .optional() .describe("Template README content (in Markdown format)"), type: z .enum(["instance", "serverless"]) .default("instance") .describe("Type of template"), channel: z .enum(["private", "community"]) .default("private") .describe("Channel for the template"), image: z .string() .nonempty() .max(500) .trim() .describe("Docker image address for instance startup"), imageAuth: z .string() .trim() .optional() .describe("ID of the container registry auth. Required only when the channel is private, or else keep it empty. You can use the `list-container-registry-auths` tool to check the ID if you're not sure."), startCommand: z .string() .max(2047) .optional() .describe("Command to run when the instance starts"), rootfsSize: z .number() .min(10) .describe("Root filesystem size (/Container disk size) in GB"), ports: z .array(z.object({ type: z.enum(["http", "udp", "tcp"]).describe("Type of exposed port"), ports: z.array( z.number() .min(1) .max(65535) .describe("Exposed port numbers, maximum of 10") ).max(10).describe("Exposed port numbers, maximum of 10"), })) .describe("Ports to expose"), env: z .array(z.object({ key: z.string().nonempty().describe("Environment variable key").max(2048), value: z.string().describe("Environment variable value").max(2048), })) .optional() .describe("Environment variables"), }), }, async (params) => { const result = await novitaRequest("/template/create", "POST", params); return { content: [ { type: "text", text: JSON.stringify(result, null, 2), }, ], }; });
  • Utility function novitaRequest used by the tool handler to perform authenticated HTTP requests to the Novita AI GPU instance API endpoints.
    export async function novitaRequest( endpoint: string, method: string = "GET", body: any = null ) { // Base URL for Novita AI API const API_BASE_URL = "https://api.novita.ai/gpu-instance/openapi/v1"; // Get API key from environment variable const API_KEY = process.env.NOVITA_API_KEY; const url = `${API_BASE_URL}${endpoint}`; const headers = { Authorization: `Bearer ${API_KEY}`, "Content-Type": "application/json", }; const options: any = { method, headers, }; if (body && (method === "POST" || method === "PATCH")) { options.body = JSON.stringify(body); } try { const response = await fetch(url, options); if (!response.ok) { const errorText = await response.text(); throw new Error(`Novita AI API Error: ${response.status} - ${errorText}`); } // Some endpoints might not return JSON const contentType = response.headers.get("content-type"); if (contentType && contentType.includes("application/json")) { return await response.json(); } return { success: true, status: response.status }; } catch (error) { console.error("Error calling Novita AI API:", error); throw error; } }

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/novitalabs/novita-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server