Wikipedia Pageviews API: Trends, Spikes, Daily Views
Pricing
from $0.17 / 1,000 pageview series delivereds
Wikipedia Pageviews API: Trends, Spikes, Daily Views
Daily Wikipedia pageviews per article and language, 0-100 interest index to compare topics, trending articles and spike alerts. Google Trends alternative on official Wikimedia data. No API key, no login. MCP-ready for AI agents. Pay per result, errors free. $0.20 per 1,000 series.
$0.20 per 1,000 pageview series (pay per result, no monthly rental). Errors, missing articles and empty results are free.
Get the daily pageviews of any Wikipedia article in any language, compare topics on a 0-100 interest index (like Google Trends), list the trending articles of a wiki with their growth, and get spike alerts when a person, brand or topic suddenly draws attention.
Data comes from the official Wikimedia Pageviews API (the same numbers as pageviews.wmcloud.org). Requests are plain JSON over HTTP: no browser, no proxy, no captcha, no API key and no login. That makes this actor fast and reliable, which matters for scheduled runs that have to work every day.
What does it do?
| Mode | Input | One result = |
|---|---|---|
series (default) | article titles, fr:Titre or Wikipedia URLs | one pageview series per article and language, with stats, trend, spikes, 0-100 interest index and share of views |
trending | a language or a list of wikis | one trending article of the latest day: rank, views, growth vs previous days, new entry flag |
alerts | your watch list of articles | only the articles that are spiking now, optionally only new spikes since the last run |
What each result includes:
dailyViews:[{date, views, interestIndex}]. Days with zero views are filled in, so the series has no gaps.totalViews,averageDailyViews,medianViews,peakViews,peakDate,latestViews,latestDatetrendPct(last 7 days vs the 7 days before) andtrendDirection(rising/stable/falling)spikes(date,views,baselineViews,ratio),spikeCount,lastSpikeDate,isSpikingNowaverageInterestIndex,shareOfComparisonPct,comparisonRankto compare the topics in one runmarkdownSummary: a ready-to-use Markdown paragraph for LLMs, reports and RAGCOMPARISON.csvin the key-value store: one column per article, one row per day, ready for Google Sheets or Excel (its link is incomparisonCsvUrl)
Redirects and spelling variants are resolved for you: elon musk becomes Elon Musk. Duplicates are counted only once. With languages, the same article is fetched in other languages through Wikipedia's interlanguage links, for example Bitcoin in en, fr, de and ja, to compare interest by country.
How much does it cost?
You pay $0.0002 per series delivered (event serie-extraite):
| Volume | Price |
|---|---|
| 10 articles × 1 language | $0.002 |
| 1,000 series (e.g. 250 brands × 4 languages) | $0.20 |
| Daily alerts on 500 articles, 3 spikes found | $0.0006 (only the 3 alerts are charged) |
| Top 50 trending articles of 5 wikis, every day for 30 days | $1.50 / month |
You are not charged for articles that don't exist, inputs with no views, network errors or an alert run that finds no spike. If you set a spending limit, the run stops cleanly when it is reached.
Input example
{"mode": "series","articles": ["ChatGPT", "Bitcoin", "fr:Emmanuel Macron", "https://de.wikipedia.org/wiki/Berlin"],"language": "en","languages": [],"lastDays": 90,"granularity": "daily","agent": "user","access": "all-access","spikeMultiplier": 3,"minSpikeViews": 1000}
Other options: startDate / endDate (YYYY-MM-DD, data starts 2015-07-01), granularity: "monthly", agent (user = humans only, all-agents, automated, spider), access (desktop, mobile-web, mobile-app), project (wiktionary, wikivoyage...), includeDailyViews, alertLookbackDays, topN, trendingCompareDays, trendingDate, sortBy (views / growth), onlySpikes, onlyNew, stateName, maxResults.
Output example (series mode, shortened)
{"mode": "series","article": "ChatGPT","title": "ChatGPT","language": "en","project": "en.wikipedia","url": "https://en.wikipedia.org/wiki/ChatGPT","input": "ChatGPT","redirected": false,"granularity": "daily","agent": "user","access": "all-access","startDate": "2026-08-02","endDate": "2026-10-31","periods": 91,"totalViews": 3412870,"averageDailyViews": 37504.1,"medianViews": 35880,"peakViews": 51583,"peakDate": "2026-10-01","latestViews": 33120,"latestDate": "2026-10-31","trendPct": -6.4,"trendDirection": "stable","spikes": [],"spikeCount": 0,"lastSpikeDate": null,"isSpikingNow": false,"comparisonRank": 1,"shareOfComparisonPct": 58.2,"averageInterestIndex": 72.7,"comparisonCsvUrl": "https://api.apify.com/v2/key-value-stores/<storeId>/records/COMPARISON.csv","dailyViews": [{ "date": "2026-10-01", "views": 51583, "interestIndex": 100 },{ "date": "2026-10-02", "views": 39904, "interestIndex": 77 },{ "date": "2026-10-03", "views": 32048, "interestIndex": 62 }],"markdownSummary": "**ChatGPT** (en.wikipedia) - 2026-08-02 to 2026-10-31 (91 days)\n- Total views: 3,412,870; ...","scrapedAt": "2026-11-01T06:00:12+00:00"}
Trending results also have rank, previousAverageViews, previousIsEstimate, growthPct and isNewEntry.
Use cases
1. Brand manager: weekly share of attention vs competitors
A consumer-electronics brand manager tracks how much public attention their brand gets compared with 3 rivals, in the 4 markets they sell in.
Input:
{ "mode": "series", "articles": ["Samsung Galaxy S26", "iPhone 18", "Google Pixel 11", "Xiaomi 16"], "languages": ["en", "fr", "de", "ja"], "lastDays": 28 }
Output (extract):
[{ "title": "iPhone 18", "language": "de", "totalViews": 412300, "shareOfComparisonPct": 21.4, "trendDirection": "rising" }]
That is 16 series for $0.0032 per run. COMPARISON.csv goes into Google Sheets every Monday, and a chart shows the share of attention by country.
2. Newsroom / PR agency: morning spike alert on a watch list
A PR agency watches 300 clients, executives and products. Every morning it only wants to see who suddenly gets attention, before journalists call.
Input (scheduled daily at 7:00):
{ "mode": "alerts", "articles": ["Satya Nadella", "Sam Altman", "fr:Bernard Arnault", "..."], "spikeMultiplier": 3, "minSpikeViews": 2000, "alertLookbackDays": 1, "onlyNew": true }
Output (extract):
[{ "title": "Sam Altman", "latestViews": 184220, "lastSpikeDate": "2026-10-31", "spikes": [{ "date": "2026-10-31", "views": 184220, "baselineViews": 21400, "ratio": 8.61 }] }]
A Make scenario posts each new alert to Slack together with its markdownSummary. Usually 0 to 5 alerts a day: a few cents a month.
3. Content / SEO team: daily breakout topics to write about
A news site's SEO editor wants the topics that are taking off today in French and Spanish, not the evergreen ones.
Input:
{ "mode": "trending", "languages": ["fr", "es"], "topN": 30, "sortBy": "growth", "onlyNew": true }
Output (extract):
[{ "rank": 7, "title": "Toussaint", "project": "fr.wikipedia", "latestViews": 98340, "previousAverageViews": 6120.4, "growthPct": 1506.8, "isNewEntry": true }]
That is 60 series, $0.012 a day. The list goes into a Google Sheet that the editorial meeting reads at 9:00.
4. Quant / research analyst: long-term interest as an alternative-data signal
An analyst pulls 5 years of monthly interest for 200 crypto assets and listed companies to correlate attention with volume.
Input:
{ "mode": "series", "articles": ["Bitcoin", "Ethereum", "Solana (blockchain platform)", "Nvidia"], "startDate": "2021-01-01", "granularity": "monthly", "includeDailyViews": true }
200 series cost $0.04. The analyst loads the results into pandas straight from the dataset API.
Call it from the API
Python:
from apify_client import ApifyClientclient = ApifyClient("<APIFY_TOKEN>")run = client.actor("<username>/wikipedia-pageviews-trends").call(run_input={"articles": ["ChatGPT", "Bitcoin"], "language": "en", "lastDays": 90,})for item in client.dataset(run["defaultDatasetId"]).iterate_items():print(item["title"], item["totalViews"], item["trendDirection"])
JavaScript:
import { ApifyClient } from 'apify-client';const client = new ApifyClient({ token: '<APIFY_TOKEN>' });const run = await client.actor('<username>/wikipedia-pageviews-trends').call({articles: ['ChatGPT', 'Bitcoin'], language: 'en', lastDays: 90,});const { items } = await client.dataset(run.defaultDatasetId).listItems();console.log(items.map((i) => [i.title, i.averageInterestIndex]));
Integrations (Make, n8n, Zapier, Google Sheets)
- Make / Zapier: use the "Run actor" module with the input JSON. Then "Get dataset items", followed by Slack, Gmail or Google Sheets "Add row".
- n8n: Apify node with the
Run actor and get datasetoperation, then filter onisSpikingNow = trueand send a Telegram or Slack message. - Google Sheets:
=IMPORTDATA("https://api.apify.com/v2/key-value-stores/<storeId>/records/COMPARISON.csv"), or the dataset export.../items?format=csv. - Schedules: create an Apify schedule (e.g. daily at 7:00) with
"mode": "alerts", "onlyNew": true.
Use with AI agents (MCP)
This actor works with the Apify MCP server, so Claude, Cursor, ChatGPT and other MCP clients can call it as a tool, for example: "Compare public interest in Taylor Swift and Beyoncé in English and Spanish over the last 6 months". Each result has a markdownSummary field that agents can quote directly, and the numeric fields can be charted without extra parsing.
FAQ
Is this the same as Google Trends? No. Google Trends measures searches; this measures how many people actually read the Wikipedia article. That is a strong public signal of interest for people, brands, events and topics, with absolute numbers rather than only relative ones, and history back to July 2015. The interestIndex puts it on a familiar 0-100 scale.
How fresh is the data? Wikimedia publishes each day's numbers after the day ends (UTC), usually within 24 hours. The latest day available is normally yesterday.
Why is my article "not found"? The title must exist on the chosen wiki. Use the exact title, a lang: prefix or paste the URL. Redirects are followed automatically. You are not charged for missing articles; they are listed in the SUMMARY record.
Are bots counted? Not by default: agent: "user" excludes identified crawlers and automated traffic. Trending lists can still contain a few bot-inflated pages (see Limits).
Do I need a Wikimedia account or key? No. The data is public and the actor only reads it.
Limits
- Data starts on 2015-07-01 and is daily or monthly, never hourly.
- The trending list is published by Wikimedia (top 1,000 per day per wiki). An article missing from a previous day's top 1,000 is estimated at that day's cut-off (
previousIsEstimate: true), so its growth is a lower bound. - Some trending pages are inflated by undetected automated traffic. Check pages with suspicious spikes before you publish.
- Pageviews measure readers of one article title. A topic split across several articles is not merged.
- Very new or tiny wikis may have no published data for the latest day.