Subscriptions
Four plans, each a superset of the last. Pick by how much of the data you need and how many things you're watching — every plan includes the full REST API and MCP server.
One thing kept current costs about 750 QU a month.
Because the data refreshes hourly, polling one query every hour is 24 calls a day — roughly 750 a month. Polling faster returns the same answer. Divide a plan's QU by 750 to get the number of things you can keep current.
The default rate. Every /api/v2 call — events, stories, entities, summaries, open feeds — and every MCP tool call counts as one Query Unit.
Keeping one entity, topic or query current at our hourly refresh cadence costs about 750 QU a month. A daily digest of the same query costs about 30.
Tool calls fired by the in-app agent bill at 5× to cover model cost. Briefs never draw down QU — each plan includes a monthly Brief count instead.
Evaluate the API
Build & prototype
Cover a beat
Talk to us
Enterprise, procurement, academic & research- Everything in Intelligence, with custom QU and request-rate limits
- Point-in-time (as-of) data — replay the database as it was known on any past date, with no look-ahead
- Uncapped alerts and scheduled Brief delivery, plus priority support
- Seat counts and custom entitlements set per contract
- Bulk / warehouse data delivery, SSO/SCIM and a formal SLA — talk to us (roadmap)
Free
Organised the way the API is.
These are the same four groups the API Arena uses, in the same order — what we produce, then what we ingest and fuse, then what we derive.
The event database we produce, plus the physical-asset directory we fuse and resolve
- CAMEO+ coded events with actors, geography & Goldstein severity
- ACLED-style conflict events — Battles, Protests, Riots, Explosions/Remote violence, Violence against civilians and Strategic developments, with fatalities, actor pairs and geocoded location. Independently generated using ACLED's published methodology and taxonomy; not sourced from or affiliated with ACLED. Beta
- Clustered stories with article-level evidence and cross-language coverage
- Wikipedia/SEC/GEM-resolved entities with a stable id spine
- The fused facilities directory — power, industry, ports and data centers on one keyed surface, 99% geocoded, with owners resolved to the entity spine where the source carries them
Source: GDELT Cloud (proprietary); facilities fused from Global Energy Monitor, NGA World Port Index, Epoch AI
Derived media measurement — also produced by us, not resold
- Cross-language tone, filterable and comparable across 24 source languages
- Share of Voice vs a custom peer set (denominator-explicit)
- Sentiment & attention timelines per entity
Source: GDELT GEG + the GDELT Cloud tone pipeline (proprietary scoring)
Public datasets we ingest and clean — with the entity-resolved feeds fused to the same spine
- Government exposure — USAspending federal awards + FARA, keyed on UEI, fused to the spine
- SEC filings — EDGAR + XBRL with AI-extracted supplier/subsidiary relations, fused to the spine
- AI & compute — Epoch AI models, hardware & data centers, fused to the spine
- Energy assets — Global Energy Monitor power, industry & fuel-transport trackers with capacity, status and owners
- Markets & macro — FRED/ALFRED with true point-in-time vintages (no look-ahead)
- Risk & screening — consolidated US restricted-party lists + DoD 1260H, as-of, with a daily change feed
- Ownership-chain exposure — who is exposed to a designated party, and by how much, with the path
- China development finance — AidData GCDF commitments (research release, through 2021)
- Maritime — chokepoint transits, port dwell & dark-gap detection over terrestrial AIS
- Legal entities — the full GLEIF Global LEI Index, with corporate hierarchy as reported to the Global LEI System (accounting consolidation, not percentage ownership)
Source: USAspending, SEC EDGAR, Epoch AI, Global Energy Monitor, FRED/ALFRED, OFAC/BIS + 1260H, AidData, terrestrial AIS, GLEIF — cited context, not a compliance determination
Intelligence products built on top of the data
- Monitoring Briefs — what changed, why it matters, the evidence, what to watch
- The research agent, over the same API a developer calls
- Alerts on any query, evaluated as often as your plan's alert cadence allows
- Atlas indices — model-derived deviation vs a place’s own norm (GPR) and its structured condition on cross-country anchors (Posture), vintaged point-in-time. Intelligence tier
- Entity Dossier — sanctions, FARA links, federal awards, events and tone for one entity in a single cited, as-of chronology. Intelligence tier
Source: GDELT Cloud (proprietary)
Our data is updated hourly, not streamed — new events are typically queryable within about an hour of the underlying reporting (median ~40 minutes), and continue to reconcile as coverage accretes. GDELT Cloud events, stories and entities run continuously from March 2026; earlier history is being backfilled and is incomplete. Maritime AIS and screening lists accrue forward from June 2026. Rows marked Early access are fully usable but still expanding. See every source, licence, coverage window and known limitation →
Have a question? Want to trial Briefs? Contact us at