Four editions a day: ~152 feeds across 19 languages are crawled — Western mainstream and wires, Western right and left, non-Western state and independent, regional outlets on every continent, the financial press, crypto and science press, and primary market data (central banks, BLS/BEA/EIA, OFAC sanctions, Treasury, OPEC, IMF). A fast model triages every article for importance and market relevance; a deep model then clusters the top slate, rewrites headlines into neutral factual form, scores real-world impact, and compares how each event is covered bloc-by-bloc.
BURIED: an important story carried by one media bloc but absent or thin in another that would be expected to cover it — verified independently against the GDELT global news index and labeled "confirmed" or "not confirmed" with the article counts shown. REFRAMED: both sides cover it, but agency/accountability is altered (e.g. "X struck Y" vs "an explosion occurred"). Flags are never dropped or promoted by the ranking — ordering is strictly by significance.
The ⚡ edge badge is deliberately strict: a story earns it only when it is both genuinely buried (verified, not merely thin in our sample) and clearly market-moving. A widely-covered story is already priced in and scores zero edge no matter how important, and importance alone is never edge. So the edge badge means one specific thing — "the market can't see this yet, and it matters" — rather than dramatic-looking headlines. Everything else is simply briefed and ranked by real-world impact.
The share of an edition's market-relevant weight (importance × market impact) that is genuinely under-covered, scaled 0–100. 0 = everything tradable is covered across blocs; higher = more of it is buried. Bands: <25 Low, 25–49 Elevated, 50–74 High, 75+ Severe.
Every outlet is treated as equally unreliable; contested claims are attributed, never asserted; no loaded language; identical scrutiny for allies and adversaries. Coverage counts are reported as data, not as truth claims about any outlet's motives.
An AI analyst can err in clustering, translation and judgement; GDELT is a partial index; coverage asymmetry can have benign causes (time zones, editorial focus). Original sources are linked on every story — verify before acting. Nothing here is investment advice.