@postman/postman-mcp-server/dist/src/tools/searchLearningCenter.js

Postman MCP Server — connect AI agents (Claude Code, Cursor, VS Code Copilot, Gemini CLI) to your Postman collections, specifications, and environments via Model Context Protocol (MCP)

import { z } from 'zod';
import { ErrorCode } from '@modelcontextprotocol/sdk/types.js';
import { asMcpError, McpError, defineToolAnnotations } from './utils/toolHelpers.js';
import { env } from '../env.js';
export const method = 'searchLearningCenter';
export const title = 'Search the Postman Learning Center';
export const description = `Search the official Postman documentation and learning resources at https://learning.postman.com.

Use this tool when you need authoritative, up-to-date guidance on how to use Postman features — for example creating mock servers, writing tests, using environments, configuring monitors, or any "how do I…" question about the Postman product. Returns relevant documentation passages with their source URLs.

Do not use this tool to search a user's own Postman resources (collections, workspaces, specs) — use \`searchPostmanElements\` for that.`;
export const parameters = z.object({
    query: z
        .string()
        .min(1)
        .max(512)
        .describe('The search query to run against the Postman documentation (e.g. "how to create a mock server", "write a test script", "set a collection variable").'),
});
export const annotations = defineToolAnnotations({
    title: 'Search the Postman Learning Center',
    readOnlyHint: true,
    destructiveHint: false,
    idempotentHint: true,
    openWorldHint: false,
});
const REMOTE_TOOL_NAME = 'searchDocs';
function parseMcpResponseBody(body, contentType) {
    const trimmed = body.trim();
    if (contentType.includes('text/event-stream') || trimmed.startsWith('event:')) {
        const dataLines = trimmed
            .split(/\r?\n/)
            .filter((line) => line.startsWith('data:'))
            .map((line) => line.slice('data:'.length).trim());
        if (dataLines.length === 0) {
            throw new Error('Learning Center MCP server returned an empty SSE response.');
        }
        return JSON.parse(dataLines.join(''));
    }
    return JSON.parse(trimmed);
}
export async function handler(args, _extra) {
    try {
        const response = await fetch(env.POSTMAN_LEARNING_CENTER_MCP_URL, {
            method: 'POST',
            headers: {
                'content-type': 'application/json',
                accept: 'application/json, text/event-stream',
            },
            body: JSON.stringify({
                jsonrpc: '2.0',
                id: 1,
                method: 'tools/call',
                params: {
                    name: REMOTE_TOOL_NAME,
                    arguments: { query: args.query },
                },
            }),
        });
        if (!response.ok) {
            const text = await response.text().catch(() => '');
            throw new McpError(ErrorCode.InternalError, `Learning Center search failed: ${response.status} ${response.statusText}${text ? ` — ${text.slice(0, 500)}` : ''}`);
        }
        const rawBody = await response.text();
        const parsed = parseMcpResponseBody(rawBody, response.headers.get('content-type') ?? '');
        if (parsed.error) {
            throw new McpError(ErrorCode.InternalError, `Learning Center search error: ${parsed.error.message}`);
        }
        if (!parsed.result) {
            throw new McpError(ErrorCode.InternalError, 'Learning Center search returned no result.');
        }
        return parsed.result;
    }
    catch (e) {
        if (e instanceof McpError) {
            throw e;
        }
        throw asMcpError(e);
    }
}