Google News Monitor: Keyword, Topic & Local Alerts

Track Google News by keyword, by section topic or by place, in any language and country edition, and get a clean deduplicated list back. Run it once for a snapshot, or schedule it in monitor mode so it returns only articles you have not seen before, optionally pushed to your webhook.

No API key, no scraping of publisher sites: it reads Google News' own public RSS feeds.

Why this one

• Three feed types in one run. Keyword searches, section topics (BUSINESS, TECHNOLOGY, SCIENCE, ...) and local news by place, combined and deduplicated together.
• Cross-feed deduplication. The same story shows up under several queries and locales. You get one row per story, with an also_matched list showing which other queries caught it, which is exactly what you want when you monitor a brand and its competitors at once.
• Real filters. Include terms, exclude terms, publisher allow list and publisher block list. Cutting the two publishers that flood every query is usually the difference between a usable alert and an ignored one.
• Monitor mode. Scheduled runs remember what they already returned and emit only new articles, so a quiet hour costs almost nothing.
• Any edition. German-Swiss, French, Italian, US, whatever hl/gl/ceid combination you need. Local coverage by city works too.

Output

{
  "title": "Example headline about the thing you monitor",
  "source": "Reuters",
  "published": "2026-07-29T19:09:00+00:00",
  "url": "https://news.google.com/rss/articles/...",
  "snippet": "Reuters",
  "feed_kind": "search",
  "feed_value": "\"prediction markets\"",
  "locale": "en-US/US",
  "also_matched": ["topic:BUSINESS"],
  "is_new": true
}

Typical setups

Brand and competitor monitoring queries: ["\"Your Company\"", "Competitor A", "Competitor B"], monitorMode: true, plus a webhookUrl. Schedule every 15 or 30 minutes.

Swiss or German local news geoPlaces: ["Zurich","Bern"], language: "de", country: "CH", ceid: "CH:de".

Sector watch with the noise removed topics: ["TECHNOLOGY","SCIENCE"], mustContain: ["AI","semiconductor"], blockSources: ["Example Aggregator"].

Input

Field	Default	Meaning
queries	none	keyword searches (quotes and OR work)
topics	none	WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, HEALTH
geoPlaces	none	city or region names for local news
language / country / ceid	en-US / US / auto	which Google News edition to read
mustContain / mustNotContain	none	keyword filters
onlySources / blockSources	none	publisher allow and block lists
monitorMode	false	only articles not returned before
webhookUrl	none	POST new articles as JSON
maxArticles	200	cap per run

Honest limitations

• Headline, publisher, timestamp, link and the feed's own short snippet. No article bodies. That is a deliberate choice: it is what the feed provides and what can be passed on. If you need full text, fetch it yourself from the publisher under their terms.
• Article links are Google News redirect links. Google now serves a JavaScript interstitial on those, so the publisher's direct URL is not recoverable server-side. The links work normally in a browser, and the publisher name is always returned separately.
• Each feed returns roughly 70 to 100 recent items, so this is a monitoring tool for current coverage rather than a historical archive. For history, run it on a schedule and let your dataset accumulate.
• Google decides which stories match a query and which publishers appear per edition. Coverage differs between editions, which is a feature when you compare them and a caveat when you do not.

Keywords

google news api, google news scraper, news monitoring, brand monitoring, media monitoring, press alerts, keyword news alerts, local news scraper, google news rss, competitor monitoring, news webhook