Worldmonitor

The honest framing from the WorldMonitor team is that this is a public-data intelligence surface, not an open-source Palantir — and the product earns that framing on both sides. On the strengths side: the breadth is real. 57 layer types, 461 feeds and 750 attributed providers, each cited, is more surface than most paid OSINT terminals expose at any price, and the correlation engine is the actual differentiator — the market lens, the commodity lens and the AI-infrastructure lens each fuse RISK (country risk, sanctions, hotspot escalation), FLOW (ships, chokepoints, flights, cables), MACRO (rates, FX, equities, crypto, energy) and PRICE into a single view rather than three tabs. The Composite Instability Index has been through eight documented methodology revisions (v8 as of this scrape), which is the kind of public re-baselining you almost never see from a free product; the changelog even tells paying API clients when to re-baseline their score caches. Transparency extends to the failure mode: when the earlier company-enrichment endpoint kept fabricating intelligence from a domain-slug heuristic, they disabled it and shipped a real SEC ticker→CIK resolution path instead. The Forecast Scorecard blog publishes a Brier-score audit of 32 AI forecasts including overconfidence analysis — self-grading that most vendors would never risk. The July 2026 additions (aviation layer, energy security dashboard, health signals, procurement feed, UAE shelf-price tracker) show the roadmap is still widening. The weaknesses are equally specific. The interface is dense by design — this is not a simple news aggregator, and there is no guided onboarding tour or preconfigured workspace. Alert rules are yours to write; the vendor does not tune them for you. Data is public-only, so if you need proprietary or private feeds in the same view you are out of category. And the correlation edge has a ceiling: geographic attribution is still a rectangular approximation rather than full point-in-polygon border geometry, which the changelog admits, so scores near contested borders (Punggye-ri, Lublin and similar) have been revised repeatedly as the resolver improved. Where it fits: geopolitical analysts verifying a breaking event against dated, cited metrics; risk teams watching chokepoints and cable cuts; journalists and researchers who need attribution built in; macro traders correlating geopolitics with markets; AI builders who want MCP or REST access to live global signals for their own agents. Where it does not: teams that want a vendor-managed analyst desk, buyers expecting a private-data terminal, and anyone who wants a light daily read rather than a working surface. In the middle sits most of the market — and for them, the free tier is the correct first move, because it costs nothing and it is genuinely the same map the Pro layer sits on top of.