Google News Scraper — Free Headlines by Keyword, Topic & Country

Get the latest Google News headlines for any keyword or topic: headline, publisher, publisher website, publish time, rank, related coverage and the Google News link. Several queries and several countries/languages in one run; schedule it and get only new articles since the last run.

Free. No usage fee; a run only uses a little Apify platform compute from your credits.

Features

• Keywords or topics: search any phrase (operators work: "exact phrase", OR, -word, site:example.com, when:7d) or pick topics TOP_STORIES, WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, HEALTH.
• 50 countries: US, GB, DE, FR, IN, BR… The language is chosen per country automatically; override with language or per edition, e.g. CH:fr, US:es.
• Time window: past hour, 12 h, 24 h, 7 days, month, year, or your own from/to dates.
• Can't run for hours: maxRunSeconds (default 300 s) stops the run cleanly and keeps results; maxItems caps the run.
• Reliable: every feed is retried up to 5 times with backoff; one bad query or country never stops the run; each query that fails gets its own item with status and errorMessage. The run only fails if every feed failed.
• Only new since last run for alerts (Slack, Sheets, n8n, Make, Zapier).
• Flat output for CSV / Google Sheets.

Free limits: up to 5 queries/topics per run (extra ones are skipped with a warning) and 500 articles per run (hard limit). One feed returns up to 100 articles per query × country; splitByDay can return more for one query, up to the 500-per-run limit.

Input

Field	Type	Default	Description
queries	string[]	–	Keywords or phrases (max 5 incl. topics)
topics	string[]	[]	Google News topics
countries	string[]	["US"]	ISO-2 codes, optionally CC:lang
language	string	native	Force a language, e.g. en, de, en-GB
maxItemsPerQuery	integer	100	1–500 per query × country (above 100 needs splitByDay)
period	enum	7d	any, 1h, 12h, 1d, 7d, 1m, 1y (search queries only)
from / to	date	empty	Custom range YYYY-MM-DD (inclusive); overrides period
splitByDay	boolean	false	Fetch each day separately + dedupe (bounded by the 500-per-run limit)
maxItems	integer	500	Cap for the whole run, 1–500
maxRunSeconds	integer	300	Clean stop at this time; unfinished queries get status: skipped
outputFormat	enum	nested	nested or flat
onlyNewSinceLastRun	boolean	false	Skip articles returned by earlier runs of the same input
stateKey	string	auto	Custom name for that memory
proxyConfiguration	object	none	Not needed normally

No input at all = US top stories.

Output

Field	Description
dataType	googleNews
status	ok; per query: no_results, skipped, error, rateLimited, invalid_input
query / topic	What returned the article
country / language	Edition
rank	Position in the feed (1 = top)
title	Headline (publisher suffix removed); titleWithSource keeps Google's original
source / sourceUrl	Publisher name and website
publishedAt	ISO 8601
googleNewsUrl / articleId	Google News link (redirects to the publisher) and its id
related / relatedCount	Other outlets covering the same story (topic feeds); flat: relatedTitles, relatedSources
feedUrl	The exact RSS URL used
coverageNote	Set when a feed/window hit Google's ~100-article cap (results may be incomplete)
errorMessage	Only on non-ok items
scrapedAt	ISO 8601

A RUN_REPORT record in the key-value store summarises feeds, successes, failures, items and duration.

{ "dataType": "googleNews", "query": "bundesliga", "topic": null, "country": "DE", "language": "de", "rank": 1,
  "title": "FC Bayern: …", "source": "SPORT1", "sourceUrl": "https://www.example.com", "publishedAt": "2026-10-23T16:30:24.000Z",
  "googleNewsUrl": "https://news.google.com/rss/articles/<articleId>", "relatedCount": 0, "status": "ok", "errorMessage": null, "scrapedAt": "2026-10-23T17:27:33.016Z" }

Only new since last run

Article ids already returned are kept in your named key-value store google-news-state under a key derived from queries + topics + countries + language (or your stateKey); last 20,000 ids, per country.

Trend → news

Run Google Trending Now, take each term and pass them as queries.

Not affiliated with Google. You are responsible for using the data lawfully.