Google News Scraper — Search & Topic Headlines RSS API

Search Google News or pull a topic's headlines — World, Business, Technology, Sports and more — in any language and country, and get back structured articles instead of a web page. Google News Scraper is a Google News API alternative built directly on Google's own public RSS feeds: give it a search query or a topic, and get one dataset item per article, ready to filter, sort, or feed into another tool.

What can you do with it?

• Monitor a brand or topic. Track every mention of your company, a competitor, or an industry term as it appears on Google News, in the language and country edition you choose.
• Build a news digest. Pull a topic's front page (Technology, Business, Science...) on a schedule and turn it into a newsletter, a Slack digest, or a dataset for another workflow.
• Feed AI agents and NLP pipelines. Call this Actor from the Apify API, an integration, or an MCP-connected AI agent so it can look up recent news on a subject as one step in a larger workflow — summarization, sentiment analysis, research.
• Track publisher coverage. See which outlets are covering a story and when, using the source and publishedAt fields.

What data do you get?

Each article Google News returns produces one dataset item:

You can download the dataset in various formats such as JSON, HTML, CSV, or Excel. Here is real, unedited output from a run with {"queries": ["artificial intelligence"], "timeRange": "7d"}:

[
  {
    "title": "What is AI, how do apps like ChatGPT work and why are there concerns?",
    "source": "bbc.com",
    "sourceUrl": "https://www.bbc.com",
    "publishedAt": "2026-09-14T16:09:56.000Z",
    "googleNewsUrl": "https://news.google.com/rss/articles/CBMiWkFVX3lxTFBiVlVDbllHd2t2NVB2YjJQYkQwOEF6ZzhqRVhyNmNQRGlzSTkzbkVrYUtLV20wMzFrWDktM0VFQXUzZEtseC0ybTRYX0xjUjI0QkZNX0M2Z3oyZw?oc=5",
    "query": "artificial intelligence",
    "language": "en",
    "country": "US",
    "error": null
  },
  {
    "title": "Artificial intelligence now beats some of the best human forecasters",
    "source": "The Economist",
    "sourceUrl": "https://www.economist.com",
    "publishedAt": "2026-09-16T18:47:08.000Z",
    "googleNewsUrl": "https://news.google.com/rss/articles/CBMixwFBVV95cUxNOVE2WlNjSlVyUjdJZzRYT1ZTSlI1cWFHdHJYcHhMNnREaXYzRzhINE8xWHZIZmh2TTFaOGllS0lTVU5sNDctUzhuOXpNQi1BbjBwalduUkE3SllUbEp3ZTFveEVZajZwOGRSVXhDaS02bnVCMkM3OWdBZTd5STl4a2dBbjNidGhpSlQ1OXlCanZRaXpoay02TU45cXk5b2c5dndkQnRCVHdCbUxvV1ljTkNtMmY4aDFnUEFHNm5iMGdhWjB4SHNR?oc=5",
    "query": "artificial intelligence",
    "language": "en",
    "country": "US",
    "error": null
  }
]

How much does it cost?

You pay $0.002 per article delivered ($2 per 1,000). Failed queries are never charged. No start fee.

Real example: the run above (

{"queries": ["artificial intelligence"], "timeRange": "7d", "maxItemsPerQuery": 10}

Input

See the Input tab for the full configuration. The minimal input is one query or one topic:

{
  "queries": ["artificial intelligence"],
  "timeRange": "7d"
}

• queries / topics — provide at least one of either. queries are free-text searches; topics are one or more of WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, HEALTH. Duplicate entries are processed once: each list is reduced to its distinct entries first, ignoring surrounding space and letter case for queries, so "ai" and "AI" are one query, fetched and billed once, not twice.
• language / country (optional, default "en" / "US") — the language and country edition of Google News to read.
• timeRange (optional) — restrict search queries to the last "1h", "1d", "7d" or "30d". Ignored for topics.
• dateFrom / dateTo (optional, YYYY-MM-DD) — restrict search queries to a date window instead. Ignored when timeRange is set.
• maxItemsPerQuery (optional, default 50, 1 to 100) — how many articles to keep from each query's or topic's feed.
• requestDelayMs (optional, default 250) — courtesy pause before each request to Google News.

Limitations — read this first

Google News returns at most about 100 items per feed, however narrow or broad the query — maxItemsPerQuery cannot get you more than the feed itself contains.

Google News links are not resolved to the publisher's final URL. googleNewsUrl is exactly the link the feed gives you — it works fine in a browser (Google decodes and redirects it client-side there), but it is not the publisher's own URL, and this Actor does not run a browser to resolve it. A plain server-side HTTP request to that link only ever lands back on news.google.com (via an intermediate Google consent/redirect page), so there is no reliable way to get the final URL without a browser — this Actor does not attempt to fake one.

Results reflect Google News' own ranking and selection for the language/country edition you chose — this Actor does not re-rank, deduplicate across publishers, or fetch full article text.

A query or topic whose feed could not be fetched leaves no row in the dataset. Unlike this studio's other Actors, every item here is an article: there is no per-query summary row, so a query that was blocked, timed out, or returned an unreadable feed simply contributes nothing. Which ones failed is in the run log (one warning per failure, naming the query or topic); the RUN_SUMMARY record in the key-value store carries only counts — failed and a tally per error category — never the names. If you need to know exactly which entries produced nothing, compare the query field of the articles you got back against the list you submitted, or read the log. Nothing that failed is ever charged.

FAQ

Is this legal? We can't give you legal advice, but here is exactly what the Actor does: it reads Google News' public RSS feeds — a format Google publishes specifically for machines to read, the same feeds any feed-reader app fetches — and returns the headline, source and link already in them. It does not log in, does not get around any block, and does not collect personal data. You are responsible for the search terms you submit and for how you use the results.

Where does the data come from? Directly from news.google.com/rss/..., Google's own public RSS endpoints for search results and topic headlines. No scraping of the Google News web app, no undocumented API.

How do I use this from the API or an AI agent (MCP)? Call it like any other Apify Actor: via the Apify API, the JavaScript or Python client, an integration, or an MCP-connected AI agent that has access to your Apify Actors. Pass the same queries / topics input and read the results back from the default dataset.

What happens if I hit my spending cap? The run stops cleanly: it finishes the article already being processed, writes everything it has, and reports chargeLimitReached: true in the run summary. Nothing already produced is lost, and nothing beyond the cap is charged.

Changelog

0.1 — Initial release.