Google News Scraper

Google News Scraper: articles by keyword or topic, with the publisher's real link

Give it a keyword or pick one of Google News's topic feeds, and get one row per article: headline, publisher, date, snippet and a link. Optionally the full body text, and a short summary with a sentiment label if you bring your own OpenAI key.

Here is the part that decides whether you need this. Every link in a Google News feed is a redirect blob, https://news.google.com/rss/articles/CBMi..., and the publisher's actual address has to be decoded out of it. That decoding is best effort. When a link will not decode, the row still arrives with the Google link in url and urlResolved: false, so you always know which one you are holding.

📰 What Google News Scraper does

It reads a Google News feed and writes a row per article: title, source, sourceUrl, publishedAt, snippet, the decoded url, and the original googleUrl in case you want it.

Search or topic. A query takes Google operators, so tesla OR rivian, site:reuters.com and intitle:layoffs all work. A topic instead gives you the edition feed for World, Nation, Business, Technology, Entertainment, Sports, Science or Health.

Full text, if you ask. With fetchArticleText on, each resolved article is downloaded and its body pulled out, up to 20,000 characters. Plenty of publishers will not give that up to a plain request, and those rows come back with articleText: null and are not charged.

Summaries, on your key. aiSummary adds a one or two sentence summary and a sentiment label, using the OpenAI key you supply. OpenAI bills you for those calls directly. This actor adds nothing for them.

📥 What you give it

{
  "query": "openai funding",
  "freshness": "7d",
  "language": "en-US",
  "country": "US",
  "maxItems": 30,
  "resolveUrls": true
}

Send neither a query nor a topic and you get one uncharged BAD_INPUT row telling you which is missing. Same if aiSummary is on without a key.

📤 What you get back

A real row from a recent run, with the long Google link cut short:

{
  "ok": true,
  "title": "Kapolri Dorong Pekerja Tingkatkan Kompetensi dan Penguasaan Teknologi - Tribratanews Polda Metro Jaya",
  "publishedAt": "2026-09-16T07:04:36.000Z",
  "source": "Tribratanews Polda Metro Jaya",
  "sourceUrl": "https://tribratanews.metro.polri.go.id",
  "snippet": "Kapolri Dorong Pekerja Tingkatkan Kompetensi dan Penguasaan Teknologi Tribratanews Polda Metro Jaya",
  "url": "https://tribratanews.metro.polri.go.id/kapolri-dorong-pekerja-tingkatkan-kompetensi-dan-penguasaan-teknologi/",
  "urlResolved": true,
  "articleText": null,
  "googleUrl": "https://news.google.com/rss/articles/CBMirgFBVV95cUxNbUp3aDlPMFNNNG5Qa3A3QlV0WFVKVjJHLU5GWExSal9qdEcz..."
}

🧾 Reading the output

There is no sample row on this one. An empty input goes straight to BAD_INPUT.

The Overview table hides urlResolved, ok, articleText and aiSummary, which are exactly the fields you want to check. Switch the dataset to JSON to see them.

▶️ How to run it

1. Open Google News Scraper and click Try for free. 2. Put keywords into Search query, or leave it empty and pick a Topic instead. 3. Set Max articles, and Freshness if you only want recent pieces. 4. Tick Fetch full article text only if you need bodies, then click Start. 5. Download the dataset as JSON, CSV or Excel, or read it from the Apify API.

💰 How much does it cost?

$3.00 per 1,000 articles, which is $0.003 each. The same rate on every Apify plan, no volume tiers.

Resolving links, fetching article text and the AI summary all cost nothing extra here. The OpenAI calls behind aiSummary are billed to you by OpenAI, on your own key.

What is not charged: a row where you asked for the body and nothing could be extracted, every diagnostic row, and a search that finds nothing.

💡 What people use it for

• Watching a brand, a competitor's product name or a ticker across publishers, on a daily schedule.
• Building a reading list for one topic feed in one country, with real links you can open.
• Pulling headline, date and publisher into a sheet for a weekly roundup, skipping the body entirely.
• Feeding resolved article URLs into your own pipeline, since googleUrl is useless to most tools.

🚧 What it does not do

• About 100 articles per feed. That is Google's cap, not ours. Split a bigger job by keyword, by

site:, or by freshness window.

• Full text is a gamble per article. Paywalls and pages that build themselves in the browser give

a plain request nothing. Those rows are flagged and not charged, so do not budget on getting the body for every article.

• Some links never decode. You get the Google link and urlResolved: false rather than a

dropped row.

• freshness narrows a search, not a topic feed. Pick a topic and you get what that feed holds.
• No archive. Google News feeds carry what is current. There is no way to ask for last March.
• The Notion connector receives every row, including the incomplete ones. Filter on ok if you

only want complete articles in there.

• One query or one topic per run. Use a schedule or several runs for a list of keywords.

🧭 Which Google scraper do you need?

❓ Questions people ask

Do I need a Google API key? No. Nothing to apply for, nothing to log into.

Do I need an OpenAI key? Only for aiSummary. Leave that off and the actor never touches OpenAI.

Why is url sometimes still a news.google.com address? Because that particular link would not decode. urlResolved tells you, per row, and googleUrl always holds the original.

Can I search several keywords at once? Not in one run. Put your operators in a single query, or schedule a run per keyword.

Why did I get fewer than maxItems? The feed ran out. A narrow keyword with a tight freshness window often returns a handful.

Is scraping Google News legal? These are public news feeds, and the rows point back at the publishers. Apify's write-up on scraping and the law is a good starting point, and we are not lawyers.

🆘 If something breaks

Open the Issues tab on the actor page. Send the query or topic and the run ID. The errorCode on the diagnostic row usually names the problem on its own.