Article Text Extractor: Clean Text & Markdown from Any URL

Turn any article, blog post, press release or docs page into clean, structured data: the main text as plain text and Markdown, plus the title, author, publish date, site name, language, lead image, tags, categories, word count and reading time. Menus, ads, cookie banners, "related stories", footers and comments are stripped out, so what you get is the article itself.

Paste a list of URLs, or point it at a homepage, a section page, an RSS/Atom feed or a sitemap and let it find the articles for you. It is plain HTTP — no browser — so it is fast and cheap, and it reads only pages the site's robots.txt allows.

What you can use it for

• LLM and RAG pipelines: feed clean Markdown (headings, lists and tables kept) into embeddings, summaries or a chatbot knowledge base instead of raw HTML full of navigation
• News monitoring: pull the full text of every new story from the feeds you follow, on a schedule
• Content research and SEO: word counts, publish dates, authors and tags across a competitor's blog
• Archiving and datasets: build a corpus of articles with clean text and consistent metadata
• Brand and PR tracking: extract the coverage your PR team collected as URLs, ready for analysis

Why this one

• $1 per 1,000 articles, and you only pay for pages where an article was actually extracted — failed, blocked or too-short pages are free
• Text and Markdown in one run (clean HTML too, if you want it)
• Rich metadata: author, publish date, site name, language, lead image, tags and categories, read from the page's own metadata
• Discovery built in: a homepage or section page, an RSS/Atom feed or an XML sitemap (index files too) all work as a starting point; the site's own feed is used first when the page advertises one
• Token-budget friendly: optional character cap per field, optional link stripping
• Honest rows: a page that shows a bot check, a 404 or no real article says so in error instead of returning junk text

Input

Field	What it does
Article URLs	Pages to extract, one per line
Discover articles from	Homepages, section pages, RSS/Atom feeds or sitemaps
Max articles per discovery source / in total	Caps — you never pay for more than this
Plain text / Markdown / Clean HTML	Which body formats to return
Keep links / Keep tables	Formatting options for Markdown and HTML
Extraction mode	Balanced (more text) or Precision (cleanest text)
Minimum words	Pages below this are reported as "not an article" (free)
Max characters per field	Cut long bodies (0 = no limit)
Only output extracted articles	Drop failed rows from the dataset
Pages at once	1–10 in parallel
Respect robots.txt	On by default

Example input:

{
  "articleUrls": ["https://en.wikipedia.org/wiki/Web_scraping"],
  "discoverFrom": ["https://techcrunch.com/feed/", "https://www.npr.org/"],
  "maxArticlesPerSite": 20,
  "includeMarkdown": true,
  "maxCharacters": 20000
}

Output

One row per page. A real example (text shortened):

{
  "url": "https://techcrunch.com/2026/10/25/5-best-apps-track-reading-discover-new-books/",
  "finalUrl": "https://techcrunch.com/2026/10/25/5-best-apps-track-reading-discover-new-books/",
  "title": "The best apps for tracking your reading and discovering new books",
  "author": "Aisha Malik",
  "publishedAt": "2026-10-25",
  "siteName": "TechCrunch",
  "hostname": "techcrunch.com",
  "language": "en",
  "description": "There are many apps that are designed to help you track your reading and discover new books, so we compiled a list of some of the best ones.",
  "image": "https://techcrunch.com/wp-content/uploads/2022/12/Screenshot-2026-10-22-at-2.48.45-PM.png?resize=1200,831",
  "tags": [
    "apps",
    "goodreads"
  ],
  "categories": [
    "Apps"
  ],
  "wordCount": 713,
  "readingTimeMinutes": 3,
  "text": "From discovering new books and organizing your reading list to tracking your reading goal and sharing recommendations, book apps make it easier to maintain reading as a hobby. There are numerous apps on the market that p…",
  "markdown": "From discovering new books and organizing your reading list to tracking your reading goal and sharing recommendations, book apps make it easier to maintain reading as a hobby. There are numerous apps on the market that p…",
  "httpStatus": 200,
  "extractedAt": "2026-10-26T10:39:05Z",
  "error": null
}

Fields: url, finalUrl, title, author, publishedAt, siteName, hostname, language, description, image, tags, categories, wordCount, readingTimeMinutes, text, markdown, html, httpStatus, extractedAt, error.

Download as JSON, CSV, Excel or HTML, or read the dataset through the API.

How much does it cost?

$0.001 per extracted article ($1 per 1,000) plus a tiny start fee. Pages with no article (errors, bot checks, pages under the minimum word count) cost nothing. Apify's free plan credit covers several thousand articles a month.

Worked examples:

1. A daily news digest: 5 feeds × 20 stories = 100 articles a day → $0.10 a day, about $3 a month
2. A RAG knowledge base from a 3,000-page docs site sitemap: 3,000 × $0.001 = $3.00, once
3. PR coverage report: 250 article links from your agency → $0.25

Higher Apify plans get an automatic discount (Bronze −5%, Silver −10%, Gold and up −15%).

FAQ

Is this legal? The actor reads publicly available pages over plain HTTP, at a polite pace, and respects each site's robots.txt by default. It does not log in, bypass paywalls or solve bot checks. What you may do with the text depends on the site's terms and the copyright rules where you are — summaries, research and internal analysis are typical uses; republishing full articles usually needs permission.

Does it work on paywalled sites? Only on what the site shows to any visitor without logging in. Paywalled parts are not extracted.

A page came back with "probably not an article". Its main text was shorter than "Minimum words" (video pages, index pages, apps that need JavaScript). Lower the minimum if you want short pages too.

Does it run JavaScript? No — it reads the HTML the server sends, which is how most news sites and blogs deliver their articles. Single-page apps that render text only in the browser will come back short.

What languages work? Any language; the language field is filled from the page.

Can I get every article on a site? Use the site's sitemap as the discovery source and raise the caps. Feeds usually only list the latest 10–50 stories.

Integrations

Schedule it and send the dataset to Google Sheets, Airtable, Slack, a webhook, a vector database or your own app with Apify integrations, Make, Zapier, n8n, LangChain or the Apify API.

Changelog

• 0.1 (Oct 2026): first release — text + Markdown + clean HTML, metadata, feed/sitemap/page discovery, free failed rows, parallel fetching.