About

What is Trafilatura Core?

Trafilatura Core is an online playground for the trafilaturacore engine. Paste raw HTML and get cleaned HTML back: boilerplate (navigation, sidebars, footers, ads) is stripped and the main content is kept. A page-type classifier routes extraction so articles, forums, product pages, and listings are each handled appropriately.

The engine runs offline. It takes HTML in and returns cleaned HTML out — it never fetches the network, so there is no crawling and no URL input.

The name is literal: Trafilatura Core is deliberately restricted to core functionality — HTML in, cleaned HTML out. Conversion to Markdown or other formats is left to other libraries that do it better.

How it works

• Boilerplate removal in four modes: precision, balanced, recall, or keep-boilerplate.
• Page-type-aware extraction driven by a classifier that scores the page before extracting.
• Content toggles to keep or drop comments, tables, images, and links.
• Optional minification of the cleaned output.

The same clean() pipeline is available as an npm CLI and as a TypeScript library. The playground generates ready-to-run snippets for both from your current settings.

To crawl and extract live websites online, try Contextractor, which is built on this engine.

Engine

trafilaturacore is a hybrid Rust and TypeScript port of Trafilatura, with a pure-Rust page-type classifier. It is an alpha, experimental build — maintained, but not yet a stable release.

Company behind

Trafilatura Core is operated by Glueo, s.r.o., a Prague-based software development company that runs its own online services and provides custom software development.

Updated: July 11, 2026