Wikipedia Scraper — Articles, Summaries & Pageview Trends

Scrape Wikipedia article data — summaries, descriptions, extracts, thumbnails, coordinates, and monthly pageview trends — into one unified dataset using the official Wikimedia APIs. No login, no proxies, no browser. Just clean, structured knowledge data.

Search by keyword or fetch exact page titles, in any language edition. Turn on pageviews to see which topics are trending.

What you get

Every article is returned as one row (recordType: "article"):

Field	Description
pageId	Wikipedia page ID
title	article title
description	short one-line description
extract	plain-text summary (intro paragraph)
type	page type (standard / disambiguation, etc.)
lang	language edition
url	link to the article
thumbnail	lead image (when available)
latitude / longitude	coordinates (for places)
timestamp	ISO time of the last revision
pageviewsTotal	total views over the window (when Include pageview trends is on)
pageviews	per-month { timestamp, views } breakdown

Input

{
  "search": ["machine learning"],
  "titles": ["Alan Turing"],
  "lang": "en",
  "searchLimit": 10,
  "includePageviews": false,
  "pageviewsMonths": 3
}

• search — keywords to search Wikipedia.
• titles — exact page titles to fetch directly.
• lang — language edition (e.g. en, de, fr, ja).
• searchLimit — max articles per search query (1–50).
• includePageviews — also fetch monthly pageviews per article.
• pageviewsMonths — how many recent months of pageviews (1–24).

Provide at least one of search or titles. Articles are de-duplicated by their canonical title (so aliases that redirect to the same page — e.g. "USA" and "United States" — return one row).

Sample output

One dataset row (an article, run with includePageviews on):

{
  "recordType": "article",
  "pageId": "1164",
  "title": "Artificial intelligence",
  "description": "Intelligence of machines",
  "extract": "Artificial intelligence (AI) is the capability of computational systems to perform tasks typically associated with human intelligence, such as learning, reasoning, problem-solving, perception, and decision-making…",
  "type": "standard",
  "lang": "en",
  "url": "https://en.wikipedia.org/wiki/Artificial_intelligence",
  "thumbnail": null,
  "latitude": null,
  "longitude": null,
  "timestamp": "2026-08-21T23:01:29.000Z",
  "pageviewsTotal": 2518174,
  "pageviews": [
    { "timestamp": "2026050100", "views": 938727 },
    { "timestamp": "2026060100", "views": 871873 },
    { "timestamp": "2026070100", "views": 707574 }
  ]
}

(thumbnail and latitude/longitude are populated for pages that have a lead image or coordinates — e.g. places and landmarks. Pageviews cover the most recent complete months only.)

Why it's reliable & cheap

Every request hits the official public Wikimedia APIs (REST summary, Action API search, Pageviews metrics) — no headless browser, no anti-bot walls, no API key. Transient errors (rate limits, 5xx, network blips) are retried automatically, and an unreachable or unknown page is reported and skipped — one bad source never fails the whole run.

Use cases

• Trending-topic analysis — rank topics by monthly pageviews
• Research & knowledge bases — pull structured summaries at scale
• Content pipelines — enrich articles, feed an AI summarizer or RAG index
• Geo datasets — collect coordinates for places and landmarks

Content from Wikipedia, available under CC BY-SA via the Wikimedia APIs.

FAQ

Do I need an API key or login? No. It uses the official public Wikimedia APIs — no key, no account, no proxy.

Is it free to try? Yes — click Try for free and run the prefilled search (machine learning) in seconds.

What are pageview trends? Turn on Include pageviews to get each article's monthly reader counts — perfect for spotting trending topics.

Can I use other languages? Yes — set lang (e.g. de, fr, ja) to query any Wikipedia edition.

Is this allowed? Wikimedia offers these public APIs for reuse; content is CC BY-SA (attribution included in the output docs).

Pricing

Pay per result — $5 per 1,000 articles. Runs that return nothing cost nothing. No monthly fee, no proxy costs.