linkedin-scraper

Scrape public LinkedIn profile pages into structured JSON. Written in TypeScript, uses Cheerio (no jQuery, no jsdom), and ships with a CLI plus an optional MCP server for AI clients.

Public guest HTML varies by profile. Some people expose experience and education; others only projects, skills, or a top card. This library returns whatever is present and never assumes a fixed shape.

Install

npm install linkedin-scraper

Requires Node.js 20+.

Library usage

import { scrapeProfile } from "linkedin-scraper";

const profile = await scrapeProfile("https://www.linkedin.com/in/williamhgates");
console.log(profile.name, profile.headline, profile.positions);

Options:

await scrapeProfile(url, {
  timeoutMs: 20_000,
  proxyUrl: process.env.PROXY_HOST, // or set PROXY_HOST / HTTPS_PROXY
  userAgent: "...",
  acceptLanguage: "en-US,en;q=0.9",
});

Parse HTML you already fetched:

import { scrapeProfileFromHtml } from "linkedin-scraper";

const profile = scrapeProfileFromHtml(html, "https://www.linkedin.com/in/someone");

Profile shape

Core fields: publicProfileUrl, name, headline, location, pictureUrl, summary.

Arrays (often empty): positions, educations, skills, projects, honors, languages, volunteering, publications, recommendations, websites.

sectionsPresent lists the data-section values found on the page so you can see what LinkedIn exposed to guests.

CLI

npx linkedin-scraper https://www.linkedin.com/in/williamhgates

Prints JSON to stdout. Non-zero exit on errors.

MCP server

Exposes a get_profile tool over stdio for Cursor, Claude Desktop, and other MCP clients.

npx linkedin-scraper-mcp

Example Cursor / Claude Desktop config:

{
  "mcpServers": {
    "linkedin-scraper": {
      "command": "npx",
      "args": ["-y", "linkedin-scraper-mcp"]
    }
  }
}

Or after a local build:

{
  "mcpServers": {
    "linkedin-scraper": {
      "command": "node",
      "args": ["/absolute/path/to/nodejs-linkedin-scraper/dist/mcp/server.js"]
    }
  }
}

Errors

Error	When
InvalidProfileUrlError	Not a /in/{slug} LinkedIn URL
ProfileNotFoundError	HTTP 404 / unparseable empty page
RateLimitedError	HTTP 999 / 429
AuthChallengeError	Auth wall / challenge HTML
FetchError	Network / timeout
ParseError	Unexpected HTML parse failure

Testing

npm test                 # offline fixtures (CI)
npm run test:live        # hits LinkedIn (LIVE_LINKEDIN=1)
npm run test:coverage

See CONTRIBUTING.md for fixture refresh notes.

Notes & limits

• Public only — no login, cookies, or Voyager API.
• LinkedIn may return HTTP 999 or challenge pages from datacenter IPs; retries, backoff, or a proxy can help.
• Respect LinkedIn’s terms of service and applicable law. You are responsible for how you use this tool.
• Package license: MIT.

Migrating from 1.x

Version 2.0 is a full rewrite:

• Callback API removed → await scrapeProfile(url)
• jquery / jsdom removed → Cheerio + fetch
• Selectors updated for current public profile markup
• Sections are optional; check sectionsPresent