trafilatura
A Python library for web page main-content extraction that extracts the primary text content from HTML, with output formats including markdown. It is commonly used for news and blog content extraction.
Timeline (last 90 days)
On news sites and Vietnamese blogs, trafilatura 2.0 extracts main content more stably than readability-lxml and newspaper4k.
A VnExpress page's raw HTML is about 180KB, but after trafilatura extraction only about 6KB remains.
LLM-ready Markdown, when built in-house by developers, typically requires post-processing combining html2text, trafilatura, or custom rules.
All Facts (3)
On news sites and Vietnamese blogs, trafilatura 2.0 extracts main content more stably than readability-lxml and newspaper4k.
50%UnverifiedA VnExpress page's raw HTML is about 180KB, but after trafilatura extraction only about 6KB remains.
50%UnverifiedLLM-ready Markdown, when built in-house by developers, typically requires post-processing combining html2text, trafilatura, or custom rules.
50%