Wikipedia Pageviews Scraper

What this Actor scrapes

The Wikimedia REST Pageviews API answers one question well: how many times was this article viewed, on this day, by this kind of visitor? This Actor takes a list of article titles and a date range and returns one clean row per article per day (or month), with view counts split by access method (desktop, mobile app, mobile web) and agent type (human, spider, automated). This is a traffic/trend feed, not a content scraper — for article text (summary, infobox, references) see our separate wikipedia-article-scraper. The two listings target different buyer intent on purpose and don't overlap.

What we handle for you

• 🔤 Title normalization done right — spaces, slashes, and non-ASCII characters (Beyoncé, AC/DC) are underscore-normalized and percent-encoded correctly before every request.
• 🫠 One bad article never fails the run — a title with no data for the requested range is logged and skipped; every other article still ships.
• 🔁 Retries with exponential backoff on 429 / 5xx — up to 5 attempts, Retry-After honoured.
• 🧊 Clean, typed dataset rows — Pydantic-validated, one row per article per day/month, JSON / CSV / Excel export straight from the Apify Console.
• 🌐 No key, no login, no proxy tax — the Wikimedia Pageviews API is public and keyless; we don't add proxy overhead you don't need.
• 💰 Pay-Per-Event pricing — you only pay for rows that land in your dataset, plus one small warm-up charge per run.

Use cases

• SEO and content strategists sizing topic demand before committing to a content push.
• PR and media researchers tracking attention spikes around a campaign or news event.
• Trend analysts building time series of public interest in a topic, product, or public figure.
• Academics and researchers studying public interest patterns at scale.

Input

Paste this into the Apify Console, or send it as the run input over the API. Proxy settings are on by default; you rarely need to touch them.

{
  "articles": [
    "Web_scraping",
    "ChatGPT",
    "Artificial_intelligence"
  ],
  "project": "en.wikipedia",
  "startDate": "20260801",
  "endDate": "20260810",
  "access": "all-access",
  "agent": "user",
  "granularity": "daily"
}

Output

One row per result, schema-validated before it is written. Export JSON, CSV, Excel or XML from the run, or read it over the API.

articleprojectaccessagentgranularitydateviews

{
  "article": "Web_scraping",
  "project": "en.wikipedia",
  "access": "all-access",
  "agent": "user",
  "granularity": "daily",
  "date": "2026-08-01",
  "views": 1842
}

Pricing

You pay only for results that land. Cap any run with maxTotalChargeUsd. See pricing & billing for worked examples.

Limitations

• A title that redirects to a canonical page reports its own (often near-zero) view count, not the canonical article's — use canonical titles for accurate numbers.
• A 404 from Wikimedia can mean either "no data for this date range" or "this article doesn't exist" — the API doesn't distinguish, so both are logged and skipped the same way.
• The separate /top most-viewed-articles endpoint isn't covered in this version — it's a fast-follow candidate.