LinkedIn-Posts-Hunter-MCP-Server

/** * LinkedIn post content extraction */ import { Page } from 'playwright'; import type { PostResult } from '../utils/types.js'; /** * Extract post description from page */ export const extractDescription = async (page: Page): Promise<string> => { try { const description = await page.$$eval( 'div.update-components-text.relative.update-components-update-v2__commentary', (divs) => divs .map((d) => { const text = (d as HTMLElement).innerText || ''; // Collapse consecutive spaces/tabs on same line, but preserve line breaks return text.replace(/[ \t]+/g, ' ').trim(); }) .filter((t) => t.length > 0) .join('\n\n') ); if (!description || description.trim().length === 0) { return '[No description content found]'; } return description; } catch (error) { return '[Description extraction failed]'; } }; /** * Extract all post content from a LinkedIn post page */ export const extractPostContent = async ( page: Page, url: string, urn: string ): Promise<PostResult> => { // Import metadata extractors dynamically to avoid circular deps const { extractProfileImage, extractAuthorName, extractAuthorOccupation, extractPostDate, extractLikeCount, extractCommentCount } = await import('./metadata.js'); // Wait for actual post description content to appear (not just the DOM element) await page.waitForFunction( () => { const el = document.querySelector('div.update-components-text.relative.update-components-update-v2__commentary'); return el && el.textContent && el.textContent.trim().length > 10; }, { timeout: 0 } ); // Extract all content in parallel const [ description, profileImage, authorName, authorOccupation, postDate, likeCount, commentCount ] = await Promise.all([ extractDescription(page), extractProfileImage(page), extractAuthorName(page), extractAuthorOccupation(page), extractPostDate(page), extractLikeCount(page), extractCommentCount(page) ]); return { link: url, description, profileImage, authorName, authorOccupation, postDate, likeCount, commentCount }; };