Machine-Readable World State for Agents and Analysts
GDELT Cloud turns global news flow into structured Events, Stories, Entities, and signals that APIs, dashboards, and AI agents can query directly.
Wednesday, June 17, 2026 UTC
Today's Stories
Conflict Events
CAMEO+ Events
Entities Today
More than raw GDELT — a full machine layer for the world
Raw news is prose. Raw GDELT is a firehose. GDELT Cloud adds the missing machine layer: classified Events, clustered Stories, linked Entities, quantified metrics — plus structured Energy Data assets from Global Energy Monitor — all exposed through REST, MCP tool categories, and a research Agent.
Conflict Events (ACLED-style)
Battles, protests, explosions, and violence coded into structured Conflict Events
CAMEO+ Events
Political, economic, technology, infrastructure, health, information, environment, and crime domains
Story Clusters
Articles grouped into coherent multi-source news narratives
Named Entities
People & orgs extracted and linked across all event datasets
Energy Data assets
Global Energy Monitor coal/gas/oil/nuclear/renewable plants, mines, pipelines, terminals, vessels — geo, capacity, ownership, status
Quantified signals
Significance, magnitude, propagation potential, market sensitivity, confidence, fatalities, and asset capacity rollups
Agent tool categories
Not another news summarizer
The value is accumulated structure: event normalization, taxonomies, entity links, temporal history, and reusable metrics.
Built for systems, not just screens
Dashboards, risk models, alerting jobs, and agents can consume the same state without rebuilding a processing pipeline.
A dependency layer for agents
Agents need external state, memory, and evidence. GDELT Cloud gives them queryable global context instead of one-off search results.
Conflict Events
AI-coded · GDELT Cloud
Sudan conflict: Khartoum urban fighting intensifies as SAF pushes into RSF-held districts
French unions stage nationwide strike over pension reform proposals
Russian drone attacks on Ukrainian infrastructure as frontline fighting continues
CAMEO+ Events
AI-coded · GDELT Cloud
US Senate debates bipartisan immigration reform package amid party negotiations
China announces expanded trade agreements with Southeast Asian partners
NATO member states expand strategic technology controls and defense procurement
News Stories
GDELT clusters · multi-source
Sudan conflict: Khartoum urban fighting intensifies
French unions stage nationwide strike over pension reform
Russian drone attacks on Ukrainian infrastructure
Named Entities
People & orgs · cross-linked
Emmanuel Macron
French president · pension protests
Russia
State actor · Ukraine conflict
SAF
Sudan Armed Forces · Khartoum
import os
import httpx
API_KEY = os.environ["GDELT_API_KEY"]
params = {
"country": "Lebanon",
"event_family": "conflict",
"has_fatalities": True,
"sort": "significance",
}
with httpx.Client(timeout=30) as client:
response = client.get(
"https://gdeltcloud.com/api/v2/events",
params=params,
headers={
"Authorization": f"Bearer {API_KEY}",
"Accept": "application/json",
},
)
response.raise_for_status()
events = response.json()["data"]API + MCP
Machine-readable world state for APIs and agents
GDELT Cloud turns global news flow into structured Events, Stories, Entities, summaries, and geographic context — plus Energy Data asset registries from Global Energy Monitor — that software can query directly through REST APIs and Agent/MCP workflows.
Clean REST API
Use standard HTTP clients to pull normalized global event objects into products, dashboards, and pipelines.
Entitlement-Based Access
API keys, rate limits, and usage tracking are tied to plans with API access.
Agent/MCP Native
Expose the same machine-readable state to Claude, Cursor, ChatGPT, LangGraph, and custom agents.
Agent access to world state
Connect GDELT Cloud to ChatGPT, Claude, Cursor, LangGraph, or your own agent runtime. The current MCP surface is compact on purpose: agents discover a category, inspect schemas, then call nested tools for structured world state and enrichment.
Progressive Discovery Wrappers
Agents see `*_tool_list`, `*_tool_get`, and `*_tool_call` wrappers instead of a sprawling top-level tool dump.
Five Tool Categories
GDELT Cloud anchors the narrative analysis; Energy Data adds physical-asset exposure; external categories enrich with markets and corroboration.
Prompts, Resources, and Skills
Research and Brief system prompts, CAMEO resources, task skills, and Brief skills are exposed natively and through tool-only transforms.
OAuth or API Key Auth
Interactive clients can use OAuth; developer agents and background jobs can use GDELT Cloud API keys.
{
"mcp_servers": {
"gdelt-cloud": {
"url":
"https://gdelt-cloud-mcp.fastmcp.app/mcp",
"auth": "bearer",
"env": {
"GDELT_API_KEY": "gdelt_sk_..."
}
}
}
}Live Connection
Monitoring Briefs
A short, source-backed intelligence memo that answers four questions: what changed, why it matters, what evidence supports it, and what to watch next. Built from structured events, clustered stories, mapped evidence, market context, and linked entities — reads like a professional memo, with the supporting data kept close at hand.
Four questions, one memo
What changed, why it matters, what evidence backs it, and what to watch next — answered in a tight, decision-ready brief tuned to your scope.
Cited and Reproducible
Every key judgment is backed by source evidence, structured metrics, negative findings, and explicit watch thresholds.
On demand or on a schedule
Run a brief in minutes or set a recurring cadence, and share polished web briefs through private links. Briefs are metered separately and never draw down your query units.
Maps and input parameters set the stage before the reader gets into the analysis.
Benchmarks, timelines, and evidence cards support the memo without taking over the page.
Now available
Research agent for multi-surface analysis
The hosted Agent UI uses the same MCP prompt and tool model: start with GDELT Cloud structured Events, Stories, and Entities, pivot to Energy Data assets (capacity, ownership, geographic exposure), then enrich with macro-finance, prediction markets, or web research when the question spans surfaces.
Agent surfaces:
Natural Language Interface
Ask complex questions in plain English while the agent handles discovery, schemas, and tool calls.
Enriched Context
Task skills guide discovery, drilldown, evidence handling, and synthesis across structured data and enrichment tools.
GDELT Research Agent
Online · MCP prompt loaded
Assess Horn of Africa risk and tell me what changed across events, markets, and source evidence.
Running GDELT discover-and-drill, then enrichment...
Analyzing Horn of Africa — summarizing Events, drilling into Stories and Entities, then checking macro and web context…
4 Surfaces
Smart Alerts
Monitor GDELT Cloud Events and Stories continuously and get notified when specific conditions are met
Natural Language
Use natural language or our Agent to create alerts around Events, Stories, geography, and categories
Multiple Frequencies
Choose from hourly or daily delivery based on your plan
Email Notifications
Instant alerts when conditions are met, delivered to your inbox
Protest Spike Alert
HourlySouth Asia region monitoring
Story Coverage Monitor
HourlyNarrative and article evidence tracking
Geopolitical Events
DailyUS-China relationship tracking
3 Active
FAQ