GDELT News Search — global news by keyword, country, language and tone (free)

Search news articles from thousands of publishers in 65 languages and get clean JSON or CSV: title, URL, publisher domain, country, language and time seen. Filter by keyword, country, language, domain, tone and date. Also returns coverage-volume and sentiment timelines for any topic. Data comes from the free GDELT DOC API — no API key, no browser, no proxies.

Typical searches: news monitoring, brand and competitor mentions, media sentiment, news by country, Google News alternative, news API for AI agents.

How to use in 3 steps

1. Enter one or more search queries (one per line). Use quotes for exact phrases: "climate change"; use OR in brackets: (solar OR wind).
2. Optionally narrow by country, language, domain, tone and time window (15min to 3m, or exact dates).
3. Click Start. Download the dataset as JSON, CSV or Excel, or schedule the Actor (e.g. hourly) and read it through the Apify API.

What you get

Field	Example
query	"artificial intelligence" (the query you entered)
title	Article headline in its original language
url	Link to the article
seenDate	2026-10-06T08:30:00Z — when GDELT first saw it
domain	irishtimes.com
sourceCountry	Ireland — publisher's country
language	English
imageUrl	Social image URL, null when none

Timeline modes return query, metric (volumePercent or averageTone), date and value instead.

{
  "query": "climate",
  "url": "https://www.irishtimes.com/environment/climate-crisis/2026/08/20/climate-obstruction-is-stronger-than-ever-despite-growing-need-for-action/",
  "title": "New era of climate obstruction requires pretend ...",
  "seenDate": "2026-10-06T08:30:00Z",
  "domain": "irishtimes.com",
  "language": "English",
  "sourceCountry": "Ireland",
  "imageUrl": null
}

Input examples

Negative English-language news about a company from the last 7 days:

{ "queries": ["\"Acme Corp\""], "timespan": "7d", "sourceLanguages": ["English"], "maxTone": -3, "maxItemsPerQuery": 100 }

How much of global news coverage mentions a topic, per day, for a given window:

{ "queries": ["semiconductor shortage"], "mode": "volumeTimeline", "startDate": "2026-09-01", "endDate": "2026-09-30" }

Japanese publishers only:

{ "queries": ["地震 OR 台風"], "sourceCountries": ["Japan"], "timespan": "1d" }

Use cases

• Brand, competitor and executive monitoring across languages
• Media sentiment and coverage trends for a topic or country
• Feeding fresh news into AI agents, RAG pipelines and dashboards
• Research datasets of news coverage

Limits

• GDELT returns up to 250 articles per query and time window; for more, split the window or run several queries.
• GDELT keeps about the last 3 months.
• GDELT asks for one request per 5 seconds, so each query takes several seconds; a busy GDELT server can slow runs down. The Actor waits and retries automatically.
• Only metadata and links are returned, not full article text.
• Tone and country filters rely on GDELT's automatic classification.

FAQ

Is this allowed? GDELT states that all its datasets are "available for unlimited and unrestricted use for any academic, commercial, or governmental use of any kind without fee" (https://www.gdeltproject.org/about.html). Please cite GDELT when you publish results.

Do I need an API key? No.

Why not Google News? This uses a different, official open source: GDELT. It is free, covers more languages and adds tone and country filters.

What does tone mean? An average sentiment score of the article, roughly -10 (very negative) to +10 (very positive); most articles are between -5 and +5.

Why is a query rejected? GDELT needs words of at least 3 characters and balanced quotes/brackets; the error text is shown in the run log.

Pricing: Free to use — you only pay Apify's standard platform usage.