Google News Monitor - Full Text & Real URLs

What does Google News Monitor do?

Google News Monitor watches Google News for your keywords, site: searches and topic sections in any country and language, and returns every article as a clean row: headline, source, publication time, the real publisher URL decoded from the Google News link and, if you want it, the full article text, author and image. Switch on Only new articles and schedule it, and each run returns just the articles you have not received yet. It is built for brand and competitor monitoring, PR and media tracking, market news for trading or research, and news feeds for AI agents and n8n, Make or Zapier workflows.

Use cases

• Brand and competitor monitoring: track mentions of your company, products and competitors in several countries at once, with duplicates across queries merged into one row.
• PR and media clipping: collect every article from the last day or hour with the publisher's own URL, ready for a clipping report or a spreadsheet.
• Company research before a call or deal: pull recent coverage of a company, including the full text, and feed it to an LLM for a summary.
• Market and industry news: follow sectors, tickers or regulators with Google News operators such as "exact phrase", OR, -exclude and site:reuters.com.
• News datasets and alerts for AI agents: a typed output schema with a description of every field, list inputs and short runs make it easy to call from agents and automation tools.

Sample output

{
  "query": "openai",
  "matchedQueries": ["openai"],
  "title": "FTC launches broad investigation into Anthropic, OpenAI",
  "sourceName": "washingtonpost.com",
  "sourceDomain": "washingtonpost.com",
  "sourceHomepage": "https://www.washingtonpost.com",
  "publishedAt": "2026-09-30T17:44:14.000Z",
  "hoursSincePublished": 0.6,
  "originalUrl": "https://www.washingtonpost.com/technology/2026/09/30/ftc-launches-broad-investigation-into-anthropic-openai/",
  "urlStatus": "resolved",
  "googleNewsUrl": "https://news.google.com/rss/articles/CBMirAFBVV95cUxPNERSOFU3...?oc=5",
  "articleId": "CBMirAFBVV95cUxPNERSOFU3...",
  "snippet": "The probe reflects a focus by Trump administration officials on using existing laws to police AI.",
  "language": "en",
  "country": "US",
  "matchedEditions": ["US:en", "GB:en"],
  "fullText": "The Federal Trade Commission has opened a broad investigation into the safety of artificial intelligence systems...",
  "fullTextStatus": "partial",
  "author": "Ian Duncan",
  "imageUrl": "https://www.washingtonpost.com/wp-apps/imrs.php?src=...",
  "found": true,
  "scrapedAt": "2026-09-30T18:20:00.000Z"
}

Every field is described in the dataset schema of the Actor. The Articles view shows the main columns and the Full text view shows author, snippet, text and image.

How much does it cost?

You pay per article returned. Pricing depends on your Apify plan: a small fee when a run starts, then a price per article that is lower on paid plans. Decoded URLs and full text are included in the article price, with no extra events. You can set a maximum spend on a run and the Actor stops when the limit is reached. The status row of a run that finds nothing, articles dropped by the time range or site: checks, and articles already delivered in only-new mode are never charged. The Apify free plan includes monthly credit you can use to try it.

Input

• Search queries: one Google News search per line. All Google News operators work, for example "electric vehicles" -tesla, openai site:reuters.com or nvidia when:12h.
• Topic sections: TOP, WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE or HEALTH.
• Editions: countries and languages as COUNTRY:language codes, for example US:en, GB:en, DE:de, FR:fr or BR:pt-419 (UK:en is read as GB:en). Each query is read in each edition. Google silently serves another edition for a pair it does not offer (PL:en returns the US:en edition), so the Actor compares the edition Google actually served with the one you asked for, skips a mismatched edition without returning or charging its articles, and names it in the run status. If none of the requested editions exists, the run fails.
• Time range: past hour, 6 or 12 hours, day, 3 days, week, month or year.
• Maximum articles: the most articles one run returns, newest first.
• Only articles new since the last run, Decode original article URLs (on by default) and Full text, author and image (off by default).

With an empty input the Actor runs a free example for "artificial intelligence" so you can see the output.

How does URL decoding work?

Google News links point to news.google.com and hide the publisher URL in an encoded article ID. The Actor decodes each returned article through the same public endpoint the Google News website uses: one request for the article page and one decoding request per 10 articles, about 1.1 requests per article. The result goes into originalUrl. The run status reports how many links were decoded and how many decoded URLs are on the source's own domain ("Original URL resolved for 20 of 20 articles, on the source's own domain for 20 of 20"). If Google stops decoding links, or if most decoded URLs point away from the article's source, the run fails with a clear message before any article is returned or charged, instead of returning rows with missing or wrong URLs.

FAQ

How many articles can one query return?

A Google News RSS feed returns about 100 articles at most (between 70 and 110 in our tests on 30 September and 1 October 2026). When a feed reaches that cap, older matches are cut off, and the run status names the capped query and edition. Use a shorter time range, a when: operator, more specific keywords, or split the query to get complete coverage. Per-feed counts are saved in the FEED_STATS record of the run.

Why is the time range checked again?

Google News does not always respect its own date operator. On 30 September 2026 the query site:bbc.co.uk when:1d returned 8 articles published between 2011 and 2025 among its 100 results. The Actor checks the publication time of every article against the time range and against when:, after: and before: in the query (after: and before: with one day of tolerance, because Google does not state the time zone of these dates), drops the articles outside and counts them in the status. site: queries are checked the same way against the source domain and the decoded publisher URL.

Why is fullText empty or short for some articles?

Full text is read with plain HTTP, without a browser. Paywalled sites return a teaser or refuse the request, and some sites render text with scripts. fullTextStatus tells you what happened for each article (ok, partial, empty, blocked, failed), and the run status gives the totals.

How often should I run it?

For monitoring, schedule it every hour or every few hours with Only articles new since the last run on. The memory is kept per combination of queries, topics, editions and time range, so different monitors never hide each other's articles.

How long does a run take?

Reading a feed takes about a second. Decoding URLs adds about 1.1 requests per article, spaced to stay polite to Google, so 20 articles take about half a minute and 100 articles one to two minutes. Full text adds about a second per article, four publishers at a time.

Is it legal to use this data?

The Actor reads public Google News feeds and public article pages without logging in and without bypassing any access control. Article texts are protected by the publishers' copyright, and how you use them is your responsibility. This is not legal advice.

Related Actors

Other public-data Actors from the same publisher, for example company job postings with salaries from Greenhouse and Ashby, are listed on the Store profile.