Wikipedia Scraper

Search any language edition and get the full text of every matching article, with a summary, categories and image. No login and no API key.

What it does

• Full article text, not just the opening paragraph. A long article comes back complete, several thousand words of clean plain text with the markup removed.
• Summary separated out - the opening paragraph is its own field, so you can use a short version without splitting the text yourself.
• Three ways in - search a topic, name exact articles, or pull an entire category.
• Any language edition - pass a code such as en, de, fr, es, hi or ja.
• Categories, image, last edited date and a word count on every row.
• Pageviews on request - the last 30 days of views plus the daily average, which is the quickest way to rank topics by real interest.
• Skip stubs with a minimum word count.

Common uses: building a training or research corpus, topic and content research, competitor and market background, populating a knowledge base, and ranking topics by attention.

Input

Field	Type	Description
searchTerms	array	Topics to search. Every match is fetched in full.
titles	array	Specific articles by exact title.
categories	array	Every article in a category.
language	string	Language edition code. Default en.
maxResultsPerTerm	integer	Cap per term or category. Default 100.
minWordCount	integer	Skip articles shorter than this.
includePageviews	boolean	Add 30 day view counts.
contactEmail	string	Optional. See the note below.
proxyConfiguration	object	Optional. Enable to spread requests across IPs.

Example input

{
  "searchTerms": ["machine learning"],
  "language": "en",
  "maxResultsPerTerm": 300,
  "minWordCount": 100,
  "includePageviews": true
}

Output

Each item is one article.

{
  "pageId": 233488,
  "title": "Example Article",
  "language": "en",
  "summary": "The opening paragraph of the article as plain text.",
  "extract": "The complete article as plain text, with the markup removed.",
  "wordCount": 8528,
  "categories": ["Example category", "Another category"],
  "categoryCount": 20,
  "imageUrl": "https://example.com/image.jpg",
  "thumbnailUrl": "https://example.com/thumb.jpg",
  "url": "https://en.wikipedia.org/wiki/Example_Article",
  "lastEdited": "2026-07-11T09:32:04Z",
  "pageviews30d": 412553,
  "pageviewsDailyAverage": 13751.8,
  "isRedirect": false,
  "searchTerm": "machine learning"
}

Notes

• No login and no API key. Enter a topic and run.
• Articles are fetched one at a time on purpose. The service returns the complete text of only one article per request, so fetching them in batches would leave most rows with empty content.
• The service limits how fast anonymous callers may read. Large runs slow down and retry rather than failing, so they finish.
• Setting contactEmail is optional but recommended for very large runs.
• Articles are deduplicated across search terms, titles and categories, so one reached two ways is delivered once and charged once.
• pageviews30d is null unless you turn pageviews on.

To improve our actors we collect anonymized usage telemetry (run stats and input patterns). No personal account data is collected.