Labsco
MCP SERVER

Scrapeless

by scrapeless-ai ↗

Fetch a page that turns ordinary scrapers away — as HTML, as Markdown or as a screenshot — or steer a cloud browser through the clicks and waits that have to happen before the content exists at all.

Site Crawling & Scraping Infrastructure✓ Verified
Summary
Two ways to reach a page, and the cheap one covers more than you would expect.

The single-call scrape tools already handle JavaScript rendering and bot protection; the session tools are for what one call cannot do — a login, a click, a wait for something to render — and the crawl tools are for when the unit is a site rather than a page. One habit worth adopting from the project's own security note: whatever comes back is text written by a stranger, so extract the part you asked for rather than passing raw page content straight into the next prompt.

What it is

Scrapeless's own server: one-shot retrieval of a URL or a Google result set, a cloud browser session you drive step by step, and an asynchronous crawl that follows links across a site.

What you get
  • A URL returned as full HTML, as Markdown for article-shaped pages, or as a screenshot taken through residential proxies
  • Google results and Google Trends interest over time, with country and language chosen per query
  • A cloud browser session created or reused and closed when you are done, its session id carried between calls
  • In-session interaction: navigation, clicks and typing against CSS selectors, key presses, scrolling to coordinates or to an element, and moving back and forward through history
  • Waiting that is either a fixed number of milliseconds or a named element appearing, so a step can hold until the content is really there
  • The live page read back as raw HTML or as visible text, plus a snapshot of the whole structure including its resources
  • A whole site crawled asynchronously from a base URL — start the job, poll it, cancel it if it is going nowhere
  • Browser profiles reused across sessions, with cookies and local storage persisted and a session timeout you set
Requirements

A Scrapeless account and its API key, created in the dashboard under API Key Management. It runs locally over stdio with SCRAPELESS_KEY in the environment, or as a hosted endpoint at https://api.scrapeless.com/mcp with the key as a header. Anything done inside a session is selector-driven: click, type, wait-for and scroll-to each need a CSS selector, so a page you have never inspected is hard to drive blind. MIT.

Setup effort

One command plus a key — npx -y scrapeless-mcp-server, then supply credentials