GDELT CloudGDELT Cloud
Snapshot assembled 2026-09-16 19:59 UTC

Data Status

What the GDELT Cloud API can serve right now, and how it got there. The top section reports what /api/v2 will actually return, so every figure there is something you can retrieve. The pipeline section answers a different question: is coding running right now.

Coded events since
2026-03-01
Clustered stories since
2026-03-08
Current through
2026-09-16

Every measurement below is a 30-day, 7-day or 48-hour window. These three dates are the whole envelope. Records dated before them exist — a 2026 article about a historical incident writes a back-dated event — but they are individual records, not coverage of those periods. Every /api/v2 list and summary response carries the same envelope as meta.coverage.

Available via the API

Read from serving_events and serving_stories — one settled row per record, which is exactly what /api/v2/events and /api/v2/stories return. Windows are whole closed UTC days; today is excluded because it settles on the hour and is complete only in arrears.

Events · last 30 closed days
72.5k
settled and retrievable
Stories · last 30 closed days
341.6k
settled and retrievable
Events settled through
2026-09-16
last settle 8 min ago
Stories settled through
2026-09-16
last settle 20 min ago

Events available per date · last 30 closed days

08-1709-0109-15

Stories available per date · last 30 closed days

08-1709-0109-15

Core API latency · last 30 days, per day

Response time for the Core endpoints — Events, Stories, Entities and Search — measured on successful API-key requests. Rejected and failed requests are excluded. Hover any bar for that period’s exact p50 and p95.

2026-08-172026-09-16
p50 (median)p95

Core API latency · last 48 hours, per hour

2026-09-14T20:00Z2026-09-16T19:00Z
p50 (median)p95

Coding pipeline · live warehouse

The processing clock: what the coder has produced in the last 60 minutes. These numbers run ahead of what the API serves by design — newly coded events become retrievable on the next refresh.

Most recent event
2026-09-16 19:52:25 UTC
coded_at
Events coded · last 1h
288
rolling 60 min
Events coded · last 24h
5.7k
rolling, processing clock
Events coded · last 30d
75.1k
includes re-codes of older dates
Stories formed · last 24h
14.8k
content clock — time_bucket

Events coded per clock hour · last 24h

Buckets are whole UTC clock hours on coded_at. The tile above is instead a rolling 60 minutes, and coding lands in hourly batches — so between batches the rolling window can sit well below the last full hour, and just after one it can sit above.

19:0007:0019:00

Stories formed per hour · last 24h

A content clock: time_bucket is the UTC hour a cluster was formed from its articles, not the hour it was processed. The two differ by 49 min at the median and 20.9 h at p95, so this chart shows when the news happened, not when we handled it.

19:0007:0019:00

Event metric distributions

Over the settled events above, bucketed by whole closed UTC days. CAMEO+ detail metrics only count CAMEO+ events with scores; Goldstein metrics count Conflict and CAMEO+ Political events. Values are the ones materialized at settle time — the same numbers the API returns.

Last 7 settled days

26.0k events

Significance

Composite 0-1 rank combining intensity, article support, fatalities, and confidence.

26.0k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Confidence

Model confidence for the structured event record.

26.0k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Goldstein scale

Signed Conflict and CAMEO+ Political valence from -10 to 10.

13.5k
-10--8
-8--6
-6--4
-4--2
-2-0
0-2
2-4
4-6
6-8
8-10

Goldstein severity

Absolute Goldstein intensity from 0 to 10, regardless of positive or negative valence.

13.5k
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10

Magnitude

CAMEO+ detail metric from 0 to 10.

24.1k
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10

Systemic importance

CAMEO+ detail metric from 0 to 1.

24.1k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Propagation potential

CAMEO+ detail metric from 0 to 1.

24.1k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Market sensitivity

CAMEO+ detail metric from 0 to 1.

24.1k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Last 30 settled days

72.5k events

Significance

Composite 0-1 rank combining intensity, article support, fatalities, and confidence.

72.5k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Confidence

Model confidence for the structured event record.

72.5k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Goldstein scale

Signed Conflict and CAMEO+ Political valence from -10 to 10.

35.9k
-10--8
-8--6
-6--4
-4--2
-2-0
0-2
2-4
4-6
6-8
8-10

Goldstein severity

Absolute Goldstein intensity from 0 to 10, regardless of positive or negative valence.

35.9k
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10

Magnitude

CAMEO+ detail metric from 0 to 10.

67.0k
0-1
1-2
2-3
3-4
4-5
5-6
6-7
7-8
8-9
9-10

Systemic importance

CAMEO+ detail metric from 0 to 1.

67.0k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Propagation potential

CAMEO+ detail metric from 0 to 1.

67.0k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Market sensitivity

CAMEO+ detail metric from 0 to 1.

67.0k
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1

Taxonomy and coverage · last 7 closed days

Every table below counts settled records, so each figure is reproducible from a /api/v2 query over the same window. The denominators are 25,956 events and 89,188 stories.

Events by family

  • CAMEO+
    24,077
  • Conflict
    1,879

Top 15 countries

Display names as the API serves them, from the settled geo column.

  • United States
    3,327
  • India
    1,724
  • Nigeria
    643
  • Spain
    641
  • United Kingdom
    626
  • Brazil
    570
  • Russia
    554
  • China
    552
  • Mexico
    537
  • Italy
    521
  • Taiwan
    494
  • Ukraine
    445
  • Canada
    439
  • Argentina
    405
  • Bangladesh
    352

Top 15 CAMEO+ subcategories

  • Make public statement
    2,216
  • Criticize or denounce
    924
  • Police investigation or operation
    845
  • Fiscal Policy Action
    683
  • Engage in negotiation
    586
  • Production capacity change
    508
  • Leadership or governance change
    428
  • Praise or endorse
    423
  • Appeal for policy change
    399
  • Express intent to change policy
    374
  • Consult
    331
  • Consider policy option
    321
  • Make a visit
    296
  • Technology development milestone
    255
  • Transport disruption
    219

Top 10 Conflict subcategories

  • Peaceful protest
    700
  • Air/drone strike
    215
  • Other
    165
  • Attack
    135
  • Arrests
    116
  • Armed clash
    109
  • Shelling/artillery/missile attack
    91
  • Protest with intervention
    70
  • Disrupted weapons use
    54
  • Remote explosive/landmine/IED
    39

Top 15 story cluster categories

  • CAMEO+ · Political
    32,481
  • CAMEO+ · Crime
    14,029
  • Conflict · Security
    11,809
  • CAMEO+ · Economic
    7,524
  • CAMEO+ · Infrastructure
    6,670
  • CAMEO+ · Environment
    5,124
  • CAMEO+ · Corporate
    4,161
  • CAMEO+ · Technology
    2,796
  • CAMEO+ · Health
    2,699
  • CAMEO+ · Information
    952
  • CAMEO+ · Demographic
    943

Top languages · across served stories

Article-languages attached to a served Story: 160,839 over this window. This is NOT the ingest corpus below — it counts only the articles that made it into a Story you can retrieve, 2.1x fewer.

  • English
    54,068
  • Spanish
    20,595
  • Arabic
    10,420
  • Russian
    9,393
  • Greek
    5,877
  • Portuguese
    4,990
  • French
    4,453
  • Chinese
    4,370
  • Albanian
    3,610
  • Italian
    3,166
  • Ukrainian
    3,160
  • Turkish
    2,528
  • Romanian
    2,524
  • German
    1,989
  • Lithuanian
    1,874
  • Serbian
    1,856
  • Bengali
    1,804
  • Bulgarian
    1,758
  • Bosnian
    1,682
  • Korean
    1,528

Top languages · ingested article corpus

Every article we language-tagged: 340,039 over the same window, whether or not it ended up in a served Story. Whole closed UTC days — the old "last 24h" label spanned 24 to 48 hours depending on the time of day.

  • English
    120,132
  • Spanish
    42,932
  • Arabic
    19,034
  • Russian
    16,414
  • Greek
    12,515
  • Chinese
    10,769
  • Portuguese
    9,271
  • French
    9,264
  • Italian
    7,885
  • Albanian
    6,296
  • German
    5,397
  • Turkish
    4,896
  • Bengali
    4,690
  • Ukrainian
    4,450
  • Romanian
    4,398
  • Lithuanian
    4,174
  • Bosnian
    3,623
  • Serbian
    3,498
  • Bulgarian
    3,297
  • Korean
    2,913

How this data is produced: GDELT Cloud codes all structured events (CAMEO+ and Conflict) and clusters Stories ourselves, using frontier LLMs against an article stream. The structured events, story clusters, and event/story metrics above are our own product — not redistributed third-party records.

Two clocks, kept apart: a content clock (the date an event or story belongs to) and a processing clock (when we coded it). What the API serves is organised by the content clock, because that is what a query filters on; the pipeline section is on the processing clock. Numbers from the two are not interchangeable and are never mixed inside one panel.

Upstream sources: raw article feed from The GDELT Project; energy infrastructure asset records from Global Energy Monitor.