Google News Scraper

Scrape Google News search results, topic sections and Top stories for any country edition. Every article comes with the real publisher URL (decoded from the Google News redirect link), the source, and the publish time.

What you get

• Decoded publisher URLs. Google News links point to news.google.com/rss/articles/.... This actor resolves each one to the article on the publisher's site, for example https://www.axios.com/2026/10/04/reflection-open-weight-ai.
• More than 100 results per search. A single Google News search returns at most about 100 articles. Set a date range and the actor searches one day at a time, so a week can return several hundred articles for a busy topic. Articles outside your range are filtered out.
• Search operators work. site:reuters.com tariffs, "climate tech" funding, intitle:nvidia, -crypto.
• Topic sections and Top stories (World, Nation, Business, Technology, Entertainment, Sports, Science, Health) for any edition, such as US, UK, France, Germany, India or Brazil.
• Checked output. Every row is validated before it is saved (title, source, valid publish date, Google News link, publisher URL not on google.com). A STATUS record lists counts, decoding results, rate limits and any errors, so you can trust a run or see exactly why it fell short.

Input example

{
  "queries": ["nvidia"],
  "dateFrom": "2026-09-28",
  "dateTo": "2026-10-04",
  "maxItemsPerQuery": 300,
  "decodeUrls": true
}

Other options: topics, includeTopHeadlines, language (hl, e.g. en-US, fr), country (gl, e.g. US, FR), ceid (advanced), timeframe (1h, 1d, 7d, 30d, 1y), proxyConfiguration.

Output example

A real row from a test run:

{
  "query": "nvidia",
  "feedType": "search",
  "topic": null,
  "dateSlice": "2026-10-04",
  "position": 2,
  "title": "Scoop: A powerful new model from startup Reflection is set to shake up the AI race",
  "source": "Axios",
  "sourceUrl": "https://www.axios.com",
  "publishedAt": "2026-10-04T12:50:25.000Z",
  "articleUrl": "https://www.axios.com/2026/10/04/reflection-open-weight-ai",
  "urlDecoded": true,
  "googleNewsUrl": "https://news.google.com/rss/articles/CBMiakFVX3lxTE5meUx1aTFnUHltcTFXRmNIWVhPbW9Nd0lZMXdQVlRuVFVNMW5xRG1QaFpnanA4YnpicU9VTnMwQlZuNzJ4M0lYV0haWnNsZjZGZzFnRkNfZkItbkg2a3BCZ1FEbHdqbEZHc0E?oc=5",
  "guid": "CBMiakFVX3lxTE5meUx1aTFnUHltcTFXRmNIWVhPbW9Nd0lZMXdQVlRuVFVNMW5xRG1QaFpnanA4YnpicU9VTnMwQlZuNzJ4M0lYV0haWnNsZjZGZzFnRkNfZkItbkg2a3BCZ1FEbHdqbEZHc0E",
  "edition": "US:en",
  "scrapedAt": "2026-10-08T01:38:41.815Z"
}

Field	Meaning
title	Headline, with the trailing " - Source" Google adds removed
source, sourceUrl	Publisher name and homepage
publishedAt	Publish time, ISO 8601 UTC
articleUrl	Real article URL on the publisher's site (null if decoding is off or failed)
urlDecoded	Whether articleUrl was resolved
googleNewsUrl, guid	Google News link and article ID
feedType, query, topic	Which search or feed returned the row
dateSlice	The daily search that returned it, when a date range is used. Google matches days loosely, so publishedAt can be a day off; it is always inside your overall range
position	Order within that query or feed
edition	Google News edition (ceid)

Test runs

Results from our own test runs on Apify, October 2026:

• nvidia, 7-day range, max 500: 500 unique articles from 155 publisher domains, all 500 URLs decoded, 0 invalid rows, about 3.3 minutes.
• openai + tesla, 40 each: 80 articles, all decoded, 41 s.
• France edition (fr/FR), Technology + Business topics + Top stories, 25 each: 75 articles, all decoded, 28 s.
• site:reuters.com tariffs and "climate tech" funding, past 30 days, decoding off: 153 articles in 5.5 s.

Speed depends on the proxy and on Google; decoding is the slow part.

Tips and limits

• Without a date range, Google returns at most about 100 articles per search. Each day in a date range also returns at most about 100, so very busy topics can still be capped per day.
• Decoding adds two requests per article. Turn it off for faster runs when the Google link is enough.
• Topic sections and Top stories come back as Google serves them for that edition, and the same story can appear in more than one feed. Rows are de-duplicated within each query or feed.
• If Google rate limits a run, the actor backs off and switches proxy sessions a limited number of times, then stops and reports it in STATUS. It does not solve CAPTCHAs.
• Only the headline and metadata Google News shows are collected, not full article text.

Pricing

Pay per result: $3.00 per 1,000 articles, plus $0.00005 per run start. No monthly rental.

FAQ

How do I get the real article URL from a Google News link?

Google News RSS links point to news.google.com/rss/articles/... redirects. This Actor decodes them and returns the original publisher URL in url, so you can fetch the article directly.

Is there a Google News API?

Google closed its News API years ago. This Actor works as a Google News API alternative: search by keyword, topic or country and export JSON or CSV.

How do I get more than 100 results for a query?

Google News returns about 100 items per search. Set a date range and the Actor splits it into smaller windows to collect more articles for the same query.

Can I monitor news about my brand or competitors?

Yes. Save the input as an Apify task and schedule it daily or hourly, then send new rows to Google Sheets, Slack or a webhook.