Google News Scraper

Get Google News search results and topic sections as structured JSON.

Pick an action and fill in its fields — the Actor fetches the data once and stores the answer in the dataset: a list as one item per element.

Input

action is required and selects what to fetch.

Each action uses only its own fields: fields of other actions are ignored and not sent. A key that is not a field of this Actor (e.g. a typo) fails the run. Required fields are checked before any request; an empty or whitespace-only string counts as a missing value.

Action	What it returns	Required	Optional
search	Search news	query	hl, gl, ceid, limit, offset, after, before, when
topic	Topic section	topic	hl, gl, ceid, limit, offset

Field	Type	Description
query	string	Search query, e.g. artificial intelligence.
topic	string, one of WORLD, NATION, BUSINESS, TECHNOLOGY, ENTERTAINMENT, SPORTS, SCIENCE, HEALTH	Google News topic section.
hl	string, search: must be a language tag like "en" or "en-US"	Host language, e.g. en-US ("search" requires a language tag like en or en-US).
gl	string, search: must be a 2-letter uppercase country code like "US"	Host country, e.g. US ("search" requires a 2-letter uppercase code).
ceid	string, search: must be in COUNTRY:language format like "US:en"	Country edition ID, e.g. US:en ("search" requires COUNTRY:language).
limit	integer, >= 1	Max number of results to return.
offset	integer, >= 0	Result offset for pagination, starting from 0.
after	string, must be a date in YYYY-MM-DD format	Start of the date range, YYYY-MM-DD (inclusive).
before	string, must be a date in YYYY-MM-DD format	End of the date range, YYYY-MM-DD (inclusive).
when	string, must be a number followed by h, d or y, like "1h", "7d" or "1y"	Open-ended date range: a number followed by h (hours), d (days) or y (years), e.g. 1h, 7d, 1y.

Example:

{"action":"search","query":"artificial intelligence","hl":"en-US","gl":"US","ceid":"US:en"}

Output

• search (items), topic (items): one dataset item per element of the list, as is. The rest of the answer (cursor, counters and other fields without the list) is stored in key-value store record PAGE_INFO; an answer with nothing but the list writes no record. An empty list stores no items.
• A rejected value (e.g. a query the source does not accept) or an item the source reports as missing is stored as one item { "success": false, "response": 2, "response_type": "invalid request", "message": "..." } ("response": 10, "page not found" when the source reports 404 or 410) and the run still succeeds. Only error items have the success field.
• The run fails without storing data when the input is invalid (checked before any request), the data source is unavailable or returns an unexpected answer, or it does not answer within 60 seconds.

Pricing

This Actor uses pay-per-event pricing: you pay for the Actor start and for each result, platform usage (compute) is included. The prices are on the Actor's page in Apify Store.

• A result is one dataset item with data: search, topic — each element of the list.
• Only the Actor start is charged for an error item, an empty list and a run that fails.
• With a maximum cost per run set, the Actor stores only as many items as the limit covers: the rest of the answer is not stored.

Run via API

curl -X POST \
  "https://api.apify.com/v2/acts/logical_variety~google-news/run-sync-get-dataset-items?token=<APIFY_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"action":"search","query":"artificial intelligence","hl":"en-US","gl":"US","ceid":"US:en"}'