Google News Scraper — headlines with the real article URL

Scrape Google News by search query or by topic, for any country and language, and get back clean records: headline, publisher, publication time, related coverage — and, optionally, the publisher's real article URL instead of Google's encrypted redirect.

No browser, no proxy, no API key. Runs in about a second.

The problem this solves

Every Google News feed hands you a link like this:

https://news.google.com/rss/articles/CBMi_AFBVV95cUxOWlBTLUU3b2YxdmZlMkx5VVEwamFhelN4OHNmWFNy...

That is a ~720-character encrypted redirect. You cannot tell which site it points to, you cannot deduplicate by domain, you cannot filter by publisher, and you cannot hand it to a scraper or an LLM to read the article. Most Google News actors stop there and leave you with that string.

Turn on Resolve publisher URLs and you get the real one:

https://www.reuters.com/technology/...

What you get per article

Field	Description
title	Headline, with the redundant " - Publisher" suffix removed
source	Publisher name
sourceUrl	Publisher home page
publishedAt	Publication time, ISO 8601 UTC
articleUrl	The publisher's real URL (when Resolve publisher URLs is on)
googleNewsUrl	The original Google News link
relatedCoverage	Other outlets covering the same story — headline, publisher and link
scrapedAt	When this run fetched the feed

To deduplicate between runs, use googleNewsUrl: it is stable for a given article.

relatedCoverage is filled on topic feeds, where Google groups the coverage of one story across outlets. It is empty on search feeds, because Google does not group those.

Input

Field	Default	Notes
query	empty	Supports Google News operators: site:, -exclusion, "exact phrase", OR. Leave empty to read a topic feed.
topic	TOP_STORIES	WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, HEALTH. Used only when there is no query.
country	US	Two-letter code: US, GB, AR, BR, DE, IN, …
language	en-US	en-US, es-419, pt-BR, de, fr, … It has to be a language Google News publishes for that country.
publishedWithin	any	1h, 12h, 1d, 7d, 1y. Searches only.
maxItems	50	Google caps a feed at about 100 articles. 0 means no limit.
resolveArticleUrls	false	Adds articleUrl. Costs one extra request per article and is charged separately.

Example — AI news from Argentina in the last 24 hours, with real URLs

{
  "query": "inteligencia artificial",
  "country": "AR",
  "language": "es-419",
  "publishedWithin": "1d",
  "maxItems": 50,
  "resolveArticleUrls": true
}

Example — technology headlines in Germany

{
  "topic": "TECHNOLOGY",
  "country": "DE",
  "language": "de"
}

Pricing

Pay per event, so you pay for what you get:

• Actor start — one small charge per run.
• Article scraped — per article returned.
• Publisher URL resolved — only when resolveArticleUrls is on, and only for URLs that actually came back. If Google refuses to resolve one, it is left null and you are not charged for it.

Reading a feed is cheap. Resolving a URL means fetching a page per article, which is why it is a separate, opt-in charge instead of a higher price for everyone.

Honest limits

• Google News returns about 100 articles per feed, and there is no pagination. For more coverage, run several narrower queries instead of one broad one.
• relatedCoverage only exists on topic feeds.
• Not every URL resolves. Some articles come from partners whose links Google will not translate; those keep articleUrl: null and are not charged.
• The feed is what Google publishes for that country/language edition. An unusual combination (for example an edition Google does not run) returns nothing — the run says so in the log rather than failing silently.

Notes

Built on the public Google News RSS feeds. No login, no personal data, no scraping of user profiles: only what Google News publishes openly.