MCP Elicitations Demo Server

import { z } from "zod"; import { zodToJsonSchema } from "zod-to-json-schema"; import { Server } from "@modelcontextprotocol/sdk/server/index.js"; import { requestTextSampling } from "../lib/sampling.js"; const SampleLLMSchema = z.object({ prompt: z.string().describe("The prompt to send to the LLM"), maxTokens: z .number() .default(100) .describe("Maximum number of tokens to generate"), systemPrompt: z .string() .optional() .describe("System prompt to guide the LLM's behavior"), }); export const sampleLlmTool = { name: "sampleLLM", description: "Samples from an LLM using MCP's sampling feature", inputSchema: zodToJsonSchema(SampleLLMSchema), handler: async (args: any, request: any, server: Server) => { const validatedArgs = SampleLLMSchema.parse(args); const { prompt, maxTokens, systemPrompt } = validatedArgs; const result = await requestTextSampling( prompt, systemPrompt, maxTokens, server, undefined // modelPreferences ); return { content: [ { type: "text" as const, text: `LLM sampling result: ${result.content.text}`, }, ], }; }, };