Global news-events API: the world's wire, queryable.
What is the world's press reporting right now? One call returns the latest articles matching a keyword — from outlets in any country, within the past N hours — drawn from GDELT's monitoring of worldwide news in 100+ languages. Metadata only, open terms.
The call
Keyless. Every article carries its outlet, link and timestamp, so an agent can cite coverage instead of asserting it — and the response is Ed25519-signed.
# keyless — keyword + country + recency, or argless for the global pulse
curl -X POST https://dynamicfeed.ai/v1/batch -H "Content-Type: application/json" \
-d '{"calls":[{"tool":"world_events","args":{"query":"flood","hours":6}}]}'
# MCP (keyless) — point any client at https://dynamicfeed.ai/mcp, then call:
world_events(query="earthquake", country="japan", hours=12)Sample response
{ "articles": [
{ "title": "Singapore builds future climate risks into today land-use choices",
"url": "https://english.news.cn/…", "outlet": "english.news.cn",
"published": "2026-06-12T10:45:00+00:00", "language": "English" } ],
"provenance": { "source": "GDELT Project",
"licence": "Open data — unlimited commercial use + redistribution w/ attribution" } }Why live data
A model's training data ends; the news doesn't. When an agent is asked about something happening now, the honest move is to look at what's actually being reported and cite it — outlet, link, timestamp — rather than improvise. GDELT machine-monitors worldwide news in 100+ languages every 15 minutes; this tool serves the latest matching coverage as clean metadata under GDELT's open terms. It's breadth, not curation: you see what the world's outlets published, including outlets you may want to weigh differently — which is why every row names its source.
Use it for
- News-aware agents that cite coverage instead of asserting facts
- Brand, market and topic monitoring across languages
- Crisis awareness alongside the hazard feeds (quakes, disasters, weather alerts)
- Media research over country and language slices
FAQ
Why metadata only?
Article bodies belong to their publishers. Titles, links, outlets and timestamps are what GDELT's open terms cover cleanly — and they're what an agent needs to find, weigh and cite coverage.
How fresh is it?
GDELT updates roughly every 15 minutes; the tool serves the most recent window (you choose the hours). Every article carries its published timestamp.
Can I use it commercially?
Yes — GDELT's terms allow unlimited commercial use and redistribution with citation, and that citation rides in the provenance envelope of every response.
Is it curated or fact-checked?
Neither — it's the raw breadth of world coverage, machine-monitored. Every row names its outlet so you (or your agent) apply your own source weighting. For ground-truth physical events, pair it with the hazard feeds.
What does it cost?
Keyless to try over MCP and POST /v1/batch; free REST key; paid plans from US$19/month — /#pricing.