Wikipedia Scraper

📚 Wikipedia Scraper

Turn Wikipedia into clean, structured data for RAG pipelines, AI agents and research. Fetch articles in any of Wikipedia's 300+ languages by title, URL, search, category (with subcategories), or random and featured lists, and get clean Markdown or plain text with sections, infobox key-values, summary, categories, links, images, Wikidata ID and last edit. It uses Wikipedia's official API, needs no login, and costs $1 per 1,000 articles.

✨ Why use Wikipedia Scraper?

• 🤖 Made for LLMs — Markdown with headings, lists, tables and LaTeX math; navigation boxes, edit links and citation markers removed; sections ready for chunking.
• 🧩 Structured extras — infobox as key-value pairs, Wikidata ID, short description, summary, categories, internal links, images with captions.
• 🌍 Every language — any Wikipedia edition, plus links to the same article in other languages.
• 🔎 Many ways in — titles, URLs, search, whole categories with subcategory depth, random articles, the featured article and most-read lists.
• ⚡ Fast and polite — the official MediaWiki API with batched metadata calls and a few parallel requests.
• 💸 Cheap — $1 per 1,000 articles, full content included.

🎯 Use cases

• 🧠 RAG knowledge bases — load topic areas or categories into a vector database, section by section.
• 🤖 AI agents — give an agent a tool that returns a clean article for any title.
• 📊 Research and analytics — compare coverage, length and edit recency across topics and languages.
• 🏷️ Entity enrichment — match products, people or places to Wikidata IDs, descriptions and infobox facts.
• 📰 Trend monitoring — schedule the daily most-read list with page views.

🚀 How to use

1. Add article titles or URLs, search queries or categories (or ask for random, featured or most-read articles).
2. Set the language and the limits.
3. Choose the content format (Markdown, plain text, both or none) and the extras you need.
4. Click Start, then export as JSON, CSV, Excel or HTML, or read the dataset through the API.

📥 Input

{
  "articles": ["Ada Lovelace", "https://en.wikipedia.org/wiki/Large_language_model"],
  "searchQueries": ["retrieval augmented generation"],
  "maxResultsPerSearch": 10,
  "language": "en",
  "contentFormat": "markdown"
}

📤 Output

One row per article. Articles found through several inputs are saved once.

Example (trimmed):

{
  "pageId": 974,
  "title": "Ada Lovelace",
  "language": "en",
  "url": "https://en.wikipedia.org/wiki/Ada_Lovelace",
  "description": "English mathematician (1815–1852)",
  "wikidataId": "Q7259",
  "isDisambiguation": false,
  "summary": "Augusta Ada King, Countess of Lovelace (née Byron; 10 December 1815 – 27 November 1852), also known as Ada Lovelace, was an English mathematician and writer chiefly known for work on Charles Babbage's...",
  "markdown": "**Augusta Ada King, Countess of Lovelace** (née **Byron**; 10 December 1815 – 27 November 1852), also known as **Ada Lovelace**, was an English mathematician and writer chiefly known for work on Charles Babbage's proposed mechanical general-purpose computer, the analytical engine. She was the first ...",
  "wordCount": 6480,
  "contentTruncated": false,
  "sections": [
    {
      "title": "Childhood",
      "level": 3,
      "path": "Biography > Childhood",
      "markdown": "Lord Byron expected his child to be a \"glorious boy\" and was disappointed when Lady Byron gave birth to a girl. The chil...",
      "words": 664
    },
    {
      "title": "Adult years",
      "level": 3,
      "path": "Biography > Adult years",
      "markdown": "Lovelace became close friends with her tutor Mary Somerville, who introduced her to Charles Babbage in 1833. She had a s...",
      "words": 675
    }
  ],
  "infobox": {
    "name": "The Right Honourable The Countess of Lovelace",
    "Born": "Augusta Ada Byron; 10 December 1815; London, England",
    "Died": "27 November 1852 (aged 36); Marylebone, London, England",
    "Resting place": "Church of St Mary Magdalene, Hucknall, Nottingham, England"
  },
  "images": [
    {
      "url": "https://upload.wikimedia.org/wikipedia/commons/e/e7/Miniature_of_Ada_Byron.jpg",
      "caption": "Ada Byron, aged four",
      "infobox": false
    },
    {
      "url": "https://upload.wikimedia.org/wikipedia/commons/d/de/Ada_Lovelace_child_portrait_Somerville_College.jpg",
      "caption": "Ada Byron, aged seven, by Alfred d'Orsay, 1822, Somerville College, Oxford",
      "infobox": false
    }
  ],
  "categories": [
    "Ada Lovelace",
    "1815 births",
    "1852 deaths",
    "19th-century British women mathematicians"
  ],
  "linkCount": 419,
  "links": [
    "Ada Lovelace (microarchitecture)",
    "Note G",
    "A. E. Chalon",
    "AMC (TV channel)",
    "A Prince of Lovers"
  ],
  "thumbnail": "https://upload.wikimedia.org/wikipedia/commons/4/4c/Ada_Lovelace_daguerreotype_by_Antoine_Claudet_1843_-_cropped.png?utm_source=en.wikipedia.org&utm_campaign=api&utm_content=thumbnail_unscaled",
  "imageUrl": "https://upload.wikimedia.org/wikipedia/commons/4/4c/Ada_Lovelace_daguerreotype_by_Antoine_Claudet_1843_-_cropped.png?utm_source=en.wikipedia.org&utm_campaign=api&utm_content=original",
  "lastEdited": "2026-10-15T19:38:07Z",
  "lastEditor": "Izzardus",
  "revisionId": 1380448378,
  "sizeBytes": 98137,
  "source": "article",
  "input": "Ada Lovelace",
  "scrapedAt": "2026-10-16T00:17:08Z"
}

💰 Pricing

You pay $1 per 1,000 articles ($0.001 per article), full content and all extras included, plus a small amount of platform usage. 100 articles cost about $0.10. Missing titles and invalid inputs are never charged.

❓ FAQ

Is it legal? Wikipedia's text is available under the Creative Commons Attribution-ShareAlike licence. Keep the article URL with the text and follow the licence when you republish.

Do I need a login or an API key? No. The actor uses Wikipedia's public API.

How many articles can I get? Search returns up to 10,000 articles per query; categories and random lists as many as you ask for.

Why are some infoboxes empty? Some articles have no infobox, and some languages (German, for example) keep person data in a different box.

Can I schedule it or use the API? Yes. Schedule runs in the console or start them through the API, for example to pull the most-read list every day.

⚠️ Limits

• Only the main article namespace is returned for search, categories and random lists.
• Wikipedia search returns at most 10,000 results per query; use more queries or categories to go further.
• The daily most-read list holds at most about 50 articles.
• The featured article and most-read lists exist only on languages that publish them.
• Templates that render as complex layouts (timelines, maps) are dropped from the Markdown.