API¶

>>> import etrafilatura as trafilatura
>>> html = """<html><head><title>Small tools win</title><meta name="author" content="Ada Lovelace">
... <meta property="og:site_name" content="Example Blog"><link rel="canonical" href="https://example.org/small-tools"></head>
... <body><nav><a href="/">Home</a> <a href="/about">About</a></nav>
... <article><h1>Small tools win</h1>
... <p>A tool that does one job is cheap to audit, cheap to replace, and composes with the rest through plain text.</p>
... <p>Platforms accumulate switching costs until leaving them costs more than staying; small tools never do.</p>
... <blockquote>Write programs that do one thing and do it well.</blockquote>
... <ul><li>grep searches</li><li>sort orders</li></ul>
... </article><footer>© 2026 Example Blog</footer></body></html>"""

>>> print(trafilatura.extract(html))
Small tools win
A tool that does one job is cheap to audit, cheap to replace, and composes with the rest through plain text.
Platforms accumulate switching costs until leaving them costs more than staying; small tools never do.
Write programs that do one thing and do it well.
- grep searches
- sort orders
>>> print(trafilatura.extract(html, output_format="markdown", with_metadata=True))
---
title: Small tools win
author: Ada Lovelace
url: https://example.org/small-tools
hostname: example.org
sitename: Example Blog
date: "2026-01-01"
fingerprint: "25687d829efc2e43"
---
# Small tools win
<BLANKLINE>
A tool that does one job is cheap to audit, cheap to replace, and composes with the rest through plain text.
<BLANKLINE>
Platforms accumulate switching costs until leaving them costs more than staying; small tools never do.
<BLANKLINE>
Write programs that do one thing and do it well.
<BLANKLINE>
- grep searches
- sort orders
>>> trafilatura.extract("<html><body></body></html>") is None
True

>>> document = trafilatura.bare_extraction(html, with_metadata=True)
>>> document.title, document.author, document.sitename
('Small tools win', 'Ada Lovelace', 'Example Blog')
>>> [child.tag for child in document.body]
['head', 'p', 'p', 'quote', 'list']
>>> list(document.as_dict())[:5]
['title', 'author', 'url', 'hostname', 'description']

>>> trafilatura.extract_with_metadata(html).url
'https://example.org/small-tools'

>>> metadata = trafilatura.extract_metadata(html)
>>> metadata.title, metadata.url, metadata.hostname, metadata.date
('Small tools win', 'https://example.org/small-tools', 'example.org', '2026-01-01')

>>> import tempfile, pathlib
>>> page = pathlib.Path(tempfile.mkdtemp()) / "page.html"
>>> _ = page.write_text(html, encoding="utf-8")
>>> trafilatura.extract(page) == trafilatura.extract(html)
True
>>> trafilatura.extract(str(page)) is None
True
>>> trafilatura.extract(str(page), input_format="auto") == trafilatura.extract(html)
True
>>> trafilatura.extract("https://example.org/article", input_format="auto")  # doctest: +SKIP
'…'

>>> trafilatura.fetch_url("https://example.org")  # doctest: +SKIP
'<!doctype html>…'
>>> trafilatura.fetch_url("not a url") is None
True

>>> trafilatura.load_html(html).tag
'html'
>>> body, text, length = trafilatura.baseline(html)
>>> body.tag, length == len(text)
('body', True)
>>> trafilatura.html2txt("<html><body><p>One</p><p>Two</p></body></html>")
'One Two'

>>> from etrafilatura.settings import Extractor, use_config
>>> options = Extractor(config=use_config(), output_format="xml", recall=True)
>>> options.focus, options.format
('recall', 'xml')
>>> trafilatura.extract(html, options=options).startswith("<doc")
True