Wikipedia Pageviews Scraper - Daily Views, Spikes, Top Articles

Get the daily pageviews of any Wikipedia article, in any language edition, from Wikimedia's official Analytics API. Each day comes with the change from the previous day and the ratio to the average of the 7 days before, so a jump in public interest is one column away. Run it on a schedule in monitoring mode and receive only the newly published days, or only the days that are spikes.

Typical uses: tracking interest in people, companies, tickers, films, games and events; newsroom and research alerts; inputs for prediction-market and trading models; checking what a country was reading yesterday.

Tip: pasting the article URL is the surest way to name an article. Wikipedia counts redirects and alternate names separately, so a typed title can point at a tiny side page. See "Read this before you trust a number".

What you get

1. Daily rows (output: daily) - one row per article per day:

{
  "status": "ok",
  "article": "AC/DC",
  "language": "en",
  "project": "en.wikipedia",
  "inputArticle": "AC/DC",
  "requestedTitle": "AC/DC",
  "resolvedTitle": "AC/DC",
  "titleVariantUsed": false,
  "otherVariantViews": null,
  "lowViewsHint": null,
  "rowType": "daily",
  "date": "2026-10-03",
  "views": 10874,
  "viewsPreviousDay": 5907,
  "changeFromPreviousDayPct": 84.1,
  "average7DaysBefore": 5939.1,
  "ratioTo7DayAverage": 1.83,
  "isSpike": true,
  "zeroFilled": false,
  "access": "all-access",
  "agent": "user",
  "articleUrl": "https://en.wikipedia.org/wiki/AC/DC",
  "scrapedAt": "2026-10-05T14:05:09.412Z"
}

2. Summary rows (output: summary) - one row per article: totalViews, averageDailyViews, peakDate, peakViews, latestDate, latestViews, latestRatioTo7DayAverage, last7DaysViews, previous7DaysViews, last7VsPrevious7Pct, spikeDays.

3. Most viewed articles of a day (topLanguages) - the ranked list Wikimedia publishes for a language edition, up to 300 rows per list: date, rank, article, views, articleUrl.

4. Monitoring mode (monitoringMode: true) - schedule the same articles once a day:

• without a spike threshold you get only the days published since the previous run (change: "new-day");
• with spikeMultiplier (for example 2) you get only the new days whose views are at least 2x the average of the 7 days before (change: "spike"), optionally with a floor (spikeMinViews);
• the first run saves the starting point, returns no day rows and is not charged; a later run with nothing new returns a single free note row saying so.

Input

Field	What it does
articles	Article URLs are the surest (https://de.wikipedia.org/wiki/Berlin; a URL carries its own language). Titles (Taylor Swift) also work.
language	Language code for plain titles: en, de, fr, es, ja, zh, simple... Default en.
days	Days of history ending yesterday (UTC). Default 30. Up to 300 days per run with daily rows, up to 3650 days with summary.
startDate, endDate	Exact range instead of days (UTC, YYYY-MM-DD). Data starts 2015-07-01.
output	daily or summary.
spikeMultiplier	Marks isSpike when views >= this many times the average of the 7 days before.
spikeMinViews	A spike must also have at least this many views.
monitoringMode, watchName, resetMonitoringState	Only new days / only spikes, with a saved starting point per article.
topLanguages, topDate, topLimit	Most viewed articles of a day per language edition.
access	all-access, desktop, mobile-web, mobile-app.
agent	user (human readers as classified by Wikimedia), all-agents, automated, spider.

Example - 90 days of three articles, flagging days at 2x the weekly average:

{ "articles": ["Taylor Swift", "Nvidia", "https://en.wikipedia.org/wiki/Bitcoin"], "days": 90, "spikeMultiplier": "2" }

Example - a daily spike alert:

{ "articles": ["Nvidia", "Tesla, Inc.", "Bitcoin"], "monitoringMode": true, "spikeMultiplier": "2", "spikeMinViews": 1000 }

Example - yesterday's most read articles on the English and German editions:

{ "topLanguages": ["en", "de"], "topLimit": 50 }

Read this before you trust a number

• Not real time. Wikimedia publishes each UTC day once, during the following day. Today's views are never available, and yesterday's may still be missing in the first hours after midnight UTC. On 2026-10-05 the figures for 2026-10-04 were already published at 10:18 UTC; the exact publishing hour was not measured and can vary.
• Paste the article URL when you can. Wikipedia counts a redirect or an alternate name separately from the article it points to: on 2026-10-04 Einstein had 53 views while Albert Einstein had 11,126. Titles are case-sensitive after the first letter.
• What the Actor does for typed titles. When you type a title (not a URL) it also tries the form with every word capitalised (albert einstein -> Albert_Einstein), keeps the one with more views in the range, and tells you in requestedTitle, resolvedTitle, titleVariantUsed and otherVariantViews (the total of the form it did not use). This costs one extra request and no extra charge. It does not resolve real redirects (USA -> United States): when a result averages fewer than 50 views a day, the rows carry a lowViewsHint saying it may be a redirect or a minor page.
• "No data" is not "no such article". For a title that does not exist, a wrong language code, or a range with no views, Wikimedia gives the same answer. The Actor returns a free article-not-found row that lists these possibilities and does not guess which one it is.
• Days with zero views are left out of Wikimedia's answer. Between the first and last day it did return, the Actor fills them in as views: 0 with zeroFilled: true.
• average7DaysBefore and ratioTo7DayAverage use the 7 days before the row's day (the day itself is not in the average). They are null when fewer than 7 earlier days exist.
• The most viewed lists are returned as Wikimedia publishes them, including the main page and special pages such as Special:Search.
• Wikimedia can revise published figures later; the Actor returns what the API says at the time of the run.
• In monitoring mode, if a run is more than 45 days after the previous one, only the latest 45 days are considered.

What is charged

• article-returned - one per article that returned data, whether that is 1 day or 3,650 days of rows, or one summary row.
• top-list-returned - one per most-viewed list.
• article-checked - monitoring mode only: one small charge per article that was checked and had nothing to return (no new day, or no spike). The first run for an article, which only saves the starting point, is not charged.
• Rows that explain why something was not returned (article-not-found, not-read, rate-limited, no-new-days, no-spike, baseline-saved, budget-reached, bad-input) are never charged. There is no start fee.

If the run reaches the maximum total charge you set, it stops, says so in a budget-reached row, and in monitoring mode leaves the memory untouched for the articles it did not return, so the next run picks them up.

Limits and good manners

The Actor sends one request at a time, two per second at most, with a User-Agent that identifies it, as Wikimedia's API guidelines ask. One article is one request regardless of the date range (two for a typed title that has a capitalised variant). Wikimedia's public API limit (200 requests a minute per User-Agent) is shared by everyone using this Actor; if you need high volumes, Wikimedia Enterprise is the service Wikimedia offers for that. If Wikimedia answers "too many requests" twice in a row the run stops and says which articles were not read, rather than hammering the API.

Data source and licence

Data comes from the Wikimedia Analytics API (pageviews). Wikimedia states that data provided by this API is available under the CC0 1.0 licence. This Actor is not affiliated with or endorsed by the Wikimedia Foundation.

Related Actors

• Google Trends Scraper - interest over time, by region, and related queries for search terms.
• Google Trends Trending Now - the daily trending searches by country.

Pageviews show what people read; search trends show what they looked for. Many teams track both.

Thanks for using this Actor

We build these tools for people who run them every day, and we improve them from what users tell us.

• Missing a field, or need another filter? Tell us in the Issues tab. If the data is there, we add it.
• Found a bug or a wrong value? Post the run ID in the Issues tab. Wrong data is the thing we fix first.

If this Actor saved you time, a short review helps other people find it.