Google News Scraper

Search Google News by keyword, topic, language and country. Get structured headlines, publisher details, publication dates and resolved publisher article URLs included in the base price. Export JSON, CSV or Excel without a Google account or proxy.

Input

{
    "queries": ["artificial intelligence"],
    "topics": [],
    "topStories": false,
    "language": "en",
    "country": "US",
    "maxArticlesPerQuery": 10,
    "dateFrom": "",
    "dateTo": "",
    "decodeUrls": true
}

Provide search queries, topics, or enable topStories. Up to 50 unique inputs per run; identical queries and topics are deduplicated. Topics: WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE and HEALTH. Google search operators work inside queries, including quotes, OR, site:reuters.com and when:7d. Invalid inputs produce free failed rows.

maxArticlesPerQuery defaults to 100 and supports up to 5,000 article rows per input. Search requests above 100, or with dateFrom/dateTo, split into daily windows, newest first, and deduplicate article IDs within each input. Dates are inclusive UTC dates, YYYY-MM-DD. If no start date is given for daily searches, the most recent 365 days ending at dateTo or today are searched. Explicit date ranges can be broader, subject to the work/time guard. Topic and top-stories feeds use one current RSS window; dates only filter that window. Queries matching the same article may produce one row per query.

Output

One flat row per article. Every key is always present; missing strings are empty, numbers are 0, arrays are empty.

Field	Meaning
inputIndex, input, inputType	Input position, original normalized input and search/topic/top_stories type
status, subStatus, error	Result classification and any error
position	Article position in that input, starting at 1; 0 for status rows
title	Headline with the trailing publisher suffix removed
publisher, publisherUrl	Publisher name and homepage from the RSS source
articleUrl	Resolved publisher URL, empty when not decoded
googleNewsUrl, articleId	Original Google News RSS article link and ID
publishedAt	Publication date in ISO 8601, or empty if missing
language, country	Requested Google News edition
relatedHeadlines	Headline strings from the RSS description
scrapedAt	Extraction timestamp

status	subStatus	Article charge
ok	article	Yes
partial	url_not_decoded	No
no_results	no_results	No
failed	invalid_input, http_CODE, blocked, network_error	No
failed	free_work_limit (internal cost/time guard)	No
failed	cost_limit_reached (your spending cap)	No

URL resolution is enabled by default. Missing results are retried once individually. If a URL still cannot be resolved, the article remains a free partial row. If you explicitly set decodeUrls=false, URL resolution is skipped, articleUrl stays empty and articles are charged at the same price. A later request failure preserves earlier saved articles and adds a free status row.

Pricing

Plan	Price per 1,000 article rows
FREE	$2.00
BRONZE	$1.50
SILVER	$1.20
GOLD and above	$1.00

Resolved publisher URLs are included. Actor start is $0.00005 per GB of run memory, minimum one event. Free partial/empty/failed rows have no article charge; the start event still applies. Your maximum charge limit is respected.

Measured performance and limits

On 2026-10-28, three search feeds returned HTTP 200 and 300/300 publisher URLs resolved from Apify datacenter IPs at 256 MB, without cookies or proxy. This is a measured sample, not a guarantee. RSS often caps search results at 100 per window; topic/top-story counts vary. Daily splitting cannot guarantee complete coverage if more than 100 matches fall in one day. Google chooses indexing and ranking; this Actor does not supply historical completeness or article bodies.

The publisher URL decoder relies on an undocumented Google endpoint that may change. A resolved URL can still lead to a paywall, moved page or a publisher restriction; publisher pages are not fetched. Language/country control the feed edition and do not translate headlines. Free work is bounded by a loss guard, so large empty or blocked searches may stop early with free_work_limit.

Python

import os
from apify_client import ApifyClient

client = ApifyClient(os.environ["APIFY_TOKEN"])
run = client.actor("sumitr_mardy/google-news-scraper").call(run_input={
    "queries": ["artificial intelligence"],
    "maxArticlesPerQuery": 10,
})
for article in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(article["status"], article["title"], article["articleUrl"])

cURL

curl -X POST 'https://api.apify.com/v2/acts/sumitr_mardy~google-news-scraper/run-sync-get-dataset-items' \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H 'Content-Type: application/json' \
  -d '{"queries":["artificial intelligence"],"maxArticlesPerQuery":10}'

Use it from AI agents (MCP)

Connect your client to the hosted Apify MCP server and sign in using Apify OAuth.

{"mcpServers":{"google-news":{"url":"https://mcp.apify.com/?tools=sumitr_mardy/google-news-scraper"}}}

For clients using an API token:

{"mcpServers":{"google-news":{"url":"https://mcp.apify.com/?tools=sumitr_mardy/google-news-scraper","headers":{"Authorization":"Bearer YOUR_APIFY_TOKEN"}}}}

For local stdio clients:

{"mcpServers":{"google-news":{"command":"npx","args":["-y","@apify/actors-mcp-server@latest","--tools","sumitr_mardy/google-news-scraper"],"env":{"APIFY_TOKEN":"YOUR_APIFY_TOKEN"}}}}

Start with 5–10 articles, inspect status/subStatus, and only use articleUrl when it is present. Use quoted brand queries for mention research; topic feeds are current headlines rather than historical search. Credentials belong in client settings, never in a repository.

This Actor is independent and is not affiliated with or endorsed by Google.