Wikipedia Article Lookup by URL or Keyword

Wikipedia exposes far more than an article's opening paragraph: a one-line description, the categories it belongs to, its images, its place on the map, the Wikidata entity behind it, its edit history, and the language editions it exists in. This actor collects that structure for articles you name by URL, exact title or keyword search — or for every page inside a Wikipedia category — and can add the full plain-text body, section outline, links, citation count, and pageview history.

Everything comes from Wikimedia's public APIs. There is no login, token, cookie or proxy setup.

Ways to feed it

• Paste URLs — desktop, mobile and index.php links all work, in any language edition. A URL's own domain decides its language.
• Name exact titles — one title per list item, in the language edition you choose. Redirects are followed.
• Search by keyword — Wikipedia's ranked full-text search, with the result snippet and rank attached to each record.
• Browse a category — every page in a category (or several) becomes an article lookup.
• Mix any of the above in one run. An article reached through several URLs, titles or searches is emitted once, with those inputs recorded as provenance.

What each record contains

Field	What it is
title	Canonical article title
url / canonicalUrl	Article URL and canonical URL
displayTitle	Title as Wikipedia renders it
summary	Lead-section plain text
fullText	Complete article body, when enabled
description	One-line description in the article's language
language	Language edition code
pageId / wikidataId	Wikipedia page ID and Wikidata entity ID
wordCount	Words in the best available text
categories	Category names
thumbnail / originalImage	Lead image objects with source, width, height
images	Viewable URLs of images used in the article
coordinates	latitude and longitude for geographic subjects
languageCount / availableLanguages	How many other editions exist, and their codes
lastModified / revisionId / lastEditor / lastEditComment / revisionSize	Latest revision metadata
isDisambiguation	Present and true for disambiguation pages
redirectedFrom	The input that redirected here
sections	Heading outline (title, level, anchor)
externalLinks	Outbound URLs cited or listed by the article
links	Titles of linked Wikipedia articles
referencesCount	Number of inline citations
pageviews	granularity, start, end, total, and points (date, views)
searchQuery / searchRank / searchSnippet	Search provenance when the article came from a keyword
requestedTitle / requestedTitles / requestedUrls / matchedSearchQueries	All inputs that resolved to this article
sourceCategory	Category that produced the record, when applicable
scrapedAt	Capture timestamp

Fields the source does not have for a particular article are omitted, never filled with null, "", [] or 0.

Inputs

Field	Type	Default	Purpose
articleUrls	list	–	Article URLs to look up
articleTitles	list	–	Exact titles to look up in language
searchQueries	list	–	Keyword searches
maxArticlesPerQuery	1–500	5	Results per search query
categoryNames	list	–	Categories to browse, with or without Category:
maxArticlesPerCategory	1–500	20	Members per category
language	374 codes	en	Edition for titles, searches and categories
maxItems	1–10000	100	Total record cap
includeFullContent	boolean	false	Add fullText
includeSections	boolean	false	Add sections
includeExternalLinks	boolean	false	Add externalLinks
includeLinks	boolean	false	Add links
includeReferences	boolean	false	Add referencesCount
includePageviews	boolean	false	Add pageviews
pageviewsGranularity	monthly / daily	monthly	Pageview point size
pageviewsPeriods	1–365	12	Months (max 12) or days

Older saved tasks keep working: keys such as urls, titles, pageTitles, query, searchQuery, maxResultsPerSearch, maxSearchResults, lang, includeFullArticle and fetchFullArticle are read and mapped inside the actor even though they are not shown in the form.

Examples

Look up two URLs:

{
  "articleUrls": [
    "https://en.wikipedia.org/wiki/Albert_Einstein",
    "https://de.wikipedia.org/wiki/Berlin"
  ]
}

Search, with enrichment:

{
  "searchQueries": ["quantum computing"],
  "maxArticlesPerQuery": 10,
  "language": "en",
  "includeFullContent": true,
  "includeSections": true,
  "includePageviews": true,
  "pageviewsGranularity": "monthly",
  "pageviewsPeriods": 6
}

Browse a category:

{
  "categoryNames": ["Physics"],
  "maxArticlesPerCategory": 50,
  "language": "en",
  "includeExternalLinks": true
}

Mixed run with titles, URLs and a search:

{
  "articleUrls": ["https://ja.wikipedia.org/wiki/東京"],
  "articleTitles": ["Python (programming language)", "Alan Turing"],
  "searchQueries": ["history of computing"],
  "maxItems": 25,
  "includeReferences": true
}

Data source and attribution

The actor reads Wikimedia's MediaWiki Action API, its parse API, and the Wikimedia pageview API. Wikipedia text is available under CC BY-SA; keep the source attribution when you republish it. This actor is independent and is not affiliated with or endorsed by the Wikimedia Foundation.

FAQ

Do I need a key, login or proxy? No. All four lookup paths work anonymously from an Apify container.

Which language editions are supported? The 374 language codes Wikimedia lists, from en and de to zh-min-nan and simple. Canonical and legacy codes map to the same edition. URL lookups take the edition from the URL; the language field applies to titles, searches and categories.

Why is summary shorter than fullText? summary is the lead section. fullText is the whole article and is only requested when you switch it on.

Can I get pageviews for any period? Yes — up to 12 monthly points or 365 daily points. The current partial month is included; daily data ends at the most recent day Wikimedia has published (usually two days ago).

What are the hard limits? Search returns at most 500 hits per query; categories, external links, internal links and language links are capped at 500 values per article, and images at 50. referencesCount counts inline citations but does not return the citation text. Rendered HTML and infobox key/value extraction are intentionally not part of this actor.

What happens to missing or non-article pages? Pages the API reports as missing are skipped and logged. Disambiguation pages are returned with isDisambiguation: true so you can filter them downstream.

Does it work with old tasks created before the revamp? Yes. The four original inputs keep their names, and earlier alias keys are mapped in code.