Google News Scraper

Search Google News by keyword, language, region, and date. Export article headlines, publisher links, and available public article details for monitoring and research.

Features

Results include the original news link, publisher name, publication date, and the query that found each article. When the publisher page can be read, results also include its description, image, and up to 2,000 characters of article text. Article text is read automatically from the page: it usually begins with the article's opening paragraphs, but it can include other page text such as photo captions, labels, links, short previews of related stories, or subscription notices, and for some subscription sites it may contain only such a notice. On a live-updates page it holds the page's latest updates, which can cover other topics than the listed headline.

Source-use restriction: Google News includes a personal, non-commercial-use restriction in its news feeds. This Actor does not grant permission to reuse Google News or publisher content. Check the applicable source terms and obtain any necessary permission before using or redistributing results.

How to Use

1. Enter one or more topics in Search Keywords.
2. Choose a region, language, and publication period.
3. Set Maximum Articles Per Query and run the Actor.
4. Open Run Summary to check collection status, then download the articles as JSON, CSV, or Excel.

Input example

{
  "searchQueries": ["AI automation", "robotics"],
  "region": "us",
  "language": "en",
  "dateFilter": "1d",
  "maxResults": 10
}

For Traditional Chinese news from Taiwan:

{
  "searchQueries": ["人工智慧", "半導體"],
  "region": "tw",
  "language": "zh-TW",
  "dateFilter": "1w",
  "maxResults": 10
}

Input

Field	Default	Description
searchQueries	Required	1–50 queries, up to 200 characters each. Emoji and some rare characters count as two or more toward this limit. Quotes and search operators are preserved. Repeated queries are included once.
region	us	Region preference. Examples: us, uk, tw, jp, de. Select from the supported options.
language	en	Language preference. Examples: en, zh-TW, zh-CN, ja, es. Use the exact code shown in the form.
dateFilter	1d	1h: past hour; 1d: past 24 hours; 1w: past 7 days; 1m: past 30 days; "": no date filter.
maxResults	20	Maximum saved articles per query, from 1 to 200. This is an upper limit, not a promised result count.

Region and language are preferences supplied to Google News; results may include other languages or coverage areas. If Google News has no edition for the selected language and region, it serves one of its other editions instead, and the run log names that edition; if a later search is sent to a different edition, the log says so too. Date filtering uses the publication dates supplied with the available results. Missing or future publication dates are excluded when a date filter is selected. When a date filter is selected and a query's search leaves it short of the requested number, the Actor searches that query once more within the period after every query has had its own search; the articles it adds are saved after those results. A query that already sets its own date limit, or whose wording an added term would change (such as an unclosed quote or a trailing OR), is searched only as typed. If Google News limits access, cannot be reached or answers in an unexpected form during one of these second searches or while reading an article it found, that article is left out, each query keeps the articles already found, the run makes no further second searches, and the run log says so.

Output

Field	Meaning
title	Article headline.
googleNewsUrl	Original Google News article link.
articleUrl	Confirmed publisher article link. A link does not guarantee that its contents can be accessed.
source	Publisher name supplied by the news result. May be empty if unavailable.
pubDate	Original publication date string, or an empty string when absent.
timestamp	Publication date as ISO 8601, or null when it cannot be verified.
description	News-result text. This can repeat the headline and publisher rather than summarize the article.
websiteName	Name supplied by the matching publisher page, or null.
websiteUrl	Origin of the confirmed publisher link. This is not a separately verified homepage.
imageUrl	Image address supplied by the matching publisher page, or null. Image availability is not checked separately.
enrichedDescription	Description supplied by the matching publisher page, or null.
excerpt	Publisher description when available; otherwise the news-result description.
articleContent	Available article text and counts, or null.
enrichmentTime	Time spent confirming and reading that article, in milliseconds.
guid	Article identifier supplied by Google News.
searchQuery	The query that saved this article. An article found by several queries is saved once.
region, language	The requested region and language preferences.
scrapedAt	ISO 8601 collection timestamp.

articleContent.content contains at most 2,000 Unicode code points. characterCount counts those returned code points. The legacy tokenCount field approximates CJK characters plus other words; it is not an AI model token count. The text may be a preview or an excerpt, not the full article.

Output example

Illustrative output below uses a fictional publisher and headline; it is not a claim that this article was collected:

{
  "title": "Example research announcement",
  "googleNewsUrl": "https://news.google.com/rss/articles/example_article_identifier",
  "articleUrl": "https://publisher.example/research",
  "pubDate": "Fri, 02 Oct 2026 00:00:00 GMT",
  "description": "Example research announcement Example Publisher",
  "source": "Example Publisher",
  "guid": "example_article_identifier",
  "timestamp": "2026-10-02T00:00:00.000Z",
  "websiteName": null,
  "websiteUrl": "https://publisher.example",
  "imageUrl": null,
  "enrichedDescription": null,
  "articleContent": null,
  "enrichmentTime": 1000,
  "excerpt": "Example research announcement Example Publisher",
  "searchQuery": "research",
  "region": "us",
  "language": "en",
  "scrapedAt": "2026-10-02T01:00:00.000Z"
}

Collection status

Run Summary is available in OUTPUT and SUMMARY in the run's key-value store. It reports saved and acknowledged article counts, articles with text, publisher metadata coverage, each query's outcome, and any stop reason. It is written when the run ends. A run aborted while it collects writes none; a Run Summary from an earlier ending of the same run stays as it was until the run is resurrected and ends again.

• collectionStatus: "complete" means the available results were processed or the requested per-query limit was reached. It does not mean every article on the topic was found or every publisher provided article text. Check articlesWithText for the number of results with text.
• collectionStatus: "partial" means some work could not finish. Review stopReason and queries; previously saved articles remain available.
• persistenceStatus: "uncertain" means a saved delivery could not be fully confirmed. Check the dataset before starting another run.

stopReason	Meaning
SOURCE_RESTRICTED	Google News limited access to the results.
SOURCE_UNAVAILABLE	Google News could not be reached.
SOURCE_CONTRACT	Some news information could not be verified, so collection stopped to protect result quality.
RESOURCE_LIMIT	The run reached its collection limit for time, requests, data or memory.
PAYMENT_LIMIT	The run reached its result allowance, such as the maximum charge set for the run.
PRICING_UNVERIFIED	The run allowance could not be verified, so no further articles were requested.
PERSISTENCE_UNCERTAIN	Saving could not be confirmed. Check the dataset before starting another run.
INTERRUPTED	The platform stopped the run before it finished, and the run could not continue collecting.
INTERNAL_ERROR	Collection could not finish. Contact support with the run link.

A run ends as Failed when it stopped with SOURCE_CONTRACT, PERSISTENCE_UNCERTAIN or INTERNAL_ERROR, or when it stopped for any reason before saving an article. Otherwise it ends as Succeeded; a run that stopped early then shows collectionStatus: "partial" and keeps the articles it saved. The run's status line names what was saved and why it stopped.

Each entry in queries has a status of complete, partial or skipped. A complete query ended because it reached Maximum Articles Per Query (query_limit) or had no more matching results that could be read (source_exhausted). A query whose second search within the period had not run when the run stopped is partial and keeps the articles it already has.

The Actor stops when Google News keeps limiting access or cannot be reached, source information cannot be verified, a run allowance is reached, or saving results becomes uncertain. During the second searches within the period, Google News limiting access, being out of reach or answering in an unexpected form ends only those searches; a run allowance, an abort or a saving problem still stops the run.

If the Apify platform moves a run to another server while it collects, the run continues there by itself. A run you abort keeps the articles it saved; resurrect it to continue the queries it had not finished. Resurrecting a run that ended with PERSISTENCE_UNCERTAIN can continue it when its saved articles can be checked against the dataset; a run that cannot continue stops again with PERSISTENCE_UNCERTAIN, or writes its Run Summary again and collects nothing more. A continuing run first checks the articles already in its dataset, does not save them again and does not search finished queries again; each unfinished query is searched again from the start, so its results can differ from those of a run that was never moved. An article that was being saved when the run stopped and is not in the dataset is left out when the run continues; because its saving and charge could not be confirmed, it counts toward the maximum charge, and the Run Summary shows persistenceStatus: "uncertain". Resurrecting a run that ended for any other reason writes its Run Summary again and collects nothing more. A run started by an earlier version of this Actor does not continue, and a run continues at most nine times.

FAQ

Why did I receive fewer articles than requested? Google News returns a limited set of available results. Date filtering and duplicate removal can reduce that set further; when a date filter is selected and a query falls short, a second search within that period can add more. An article found by multiple queries is saved at most once, under the first query that saves it. Run allowances and access restrictions can also stop collection early; check Run Summary.

What happens if my run is moved to another server or aborted? A run that the Apify platform moves to another server continues there by itself. A run you abort keeps the articles it saved; resurrect it to continue. Articles saved before the move or the abort are not saved again, and queries that were not finished are searched again.

Why is article text missing even though a publisher link is present? A publisher may deny access, require a subscription, send the link to a page that is not the article (such as a home or consent page), render content unavailable to this Actor, serve a page far larger or slower to read than an article page, or provide no verifiable article text. Publisher details are kept only when the page can be matched to the listed article, for example because it states the article's headline. A live-updates page whose headline has changed since it was listed can therefore give no details. The confirmed link and news-result fields can still be useful. Missing publisher information stays null.

Why does some article text mention other topics? Article text is read automatically from the publisher page. Some pages place short previews of related stories inside the article, and a live-updates page shows its latest updates, which can be about other topics than the listed headline. Open the article link to read the publisher's page as published.

Why are the results in a different language than I selected? Google News does not offer every language in every region. When it has no edition for your combination, it serves one of its other editions, and the run log says which one. Choose a language that Google News offers for that region to receive results in that language.

Does this search the full news archive or guarantee current breaking news? No. It processes the results Google News makes available for that search. Publication dates, ranking, coverage, and updates are controlled by the sources. Removing the date filter does not provide a complete historical archive.

Can I combine different languages? Queries may contain mixed-language text. Each run uses one selected language and region preference. Use separate runs for different region and language combinations.

Should I give the run more memory? Keep the default 1,024 MB. Above 1,024 MB most of the run's time allowance shrinks in proportion to the memory you choose, so a larger memory setting does not collect more articles and can end a long run sooner.

Where can I check charges? Review the current pricing shown on the Actor page and your run's usage details. Set an appropriate run allowance before starting. A zero-result or interrupted run does not establish that no platform resources were consumed.

Disclaimer

This Actor is independently developed and is not affiliated with Google or any publisher. Source availability, accuracy, completeness, and content rights are controlled by their respective providers. It does not bypass subscriptions or grant a content license. Follow applicable source terms, copyright requirements, and privacy obligations. Verify important information against the original publisher before relying on it.

Google News, news search, article metadata, publisher links, Traditional Chinese news, Claude Code, OpenClaw, Gemini, Antigravity, Codex, ChatGPT