News Radar — Phase 1

Real-time market shock detection

QuantGist groups breaking news into scored, asset-linked event clusters — so you trade on signal, not noise.

Live cluster preview

Breaking clusters, right now

Each card is a real-time event cluster: topic-packed, impact-scored, and asset-tagged the moment it clears the severity threshold.

iran-warbreaking

Iran Launches Drone Strikes on Israeli Military Bases — Oil Spikes 4.2%

Impact

0.85

Affected assets

XAU/USDWTIUSD/JPYVIX

why_it_matters✦ AI

Escalation above threshold 3 — strait blockade risk elevated. Gold and WTI are historical safe-haven proxies in this scenario; USD/JPY pricing tail risk.

trump-postsbreaking

Trump Posts: "25% Tariffs on All Chinese Goods Starting Monday"

Impact

0.84

Affected assets

USD/CNHAUD/USDSPYEEM

why_it_matters✦ AI

Unscheduled trade policy shock. CNH, AUD, and EM FX historically reprice within 2 minutes of similar posts. SPX futures already -0.6% pre-market.

oil-supplydeveloping

OPEC+ Emergency Call Signals Output Cut Extension Through Q3

Impact

0.68

Affected assets

WTIXLEBRL/USDCAD/USD

why_it_matters

Confirmation count at 2/5 — cluster still forming. WTI upside case intact if Libya outage data is confirmed in the next feed cycle.

breaking ≥ 0.80 developing 0.60–0.79fading < 0.60✦ AI — LLM-enhanced summary

How it works

Four layers that turn noise into edge

Every cluster passes through the full scoring and enrichment pipeline before it reaches your system.

7-signal impact scoring

Every cluster is scored across severity, velocity, credibility, recency, novelty, asset linkage, and confirmation count — then collapsed to a single 0–1 impact float your model can threshold directly.

Asset-linked by design

Each cluster ships with the exact tickers and instruments that historically move on that topic pack — no extra enrichment step needed.

AI-enhanced summaries

ClusterSummarizer upgrades template text to LLM-generated copy for clusters scoring ≥ 0.70 impact, giving you a trader-readable why_it_matters in every response.

Lifecycle status

Clusters progress through developing → breaking → fading → resolved. Poll the same cluster_id over time or filter by status to get only active shocks.

01Ingest

9 topic packs monitor 60+ curated RSS and API sources in parallel

02Score

ShockScorer aggregates 7 sub-signals into a single 0–1 impact float

03Deliver

Clusters above threshold appear in /v1/news/radar within 15 seconds

Code example

One call to get breaking clusters

The response ships with impact score, status, affected assets, and an AI-enhanced summary — all in one round-trip.

Python SDK

# Fetch breaking clusters with impact ≥ 0.65
from quantgist import QuantGistClient

client = QuantGistClient(api_key="qg_live_...")
clusters = client.news.radar(
    min_impact=0.65,
    status="breaking",
)

# → [ClusterResult, ClusterResult, ...]
print(clusters[0])

Response

{
  "cluster_id": "iran-war-20260528T0942Z",
  "topic":      "iran-war",
  "status":     "breaking",
  "impact":     0.85,
  "headline":   "Iran Launches Drone Strikes on Israeli Military Bases",
  "affected_assets": ["XAU/USD", "WTI", "USD/JPY", "VIX"],
  "why_it_matters": "Escalation above threshold 3 — strait blockade risk...",
  "ai_enhanced": true,
  "first_seen":  "2026-05-28T09:42:11Z",
  "updated_at":  "2026-05-28T09:58:03Z"
}

REST endpoint: GET /v1/news/radar · topic packs: GET /v1/news/topics

9 topic packs

Pre-built watchlists for the scenarios that move markets

Each topic pack is a curated set of keywords, entities, and source weights tuned for a specific shock scenario.

iran-wartrump-postsoil-supplycentral-bank-unscheduledsanctionsmiddle-east-riskcyberattackceasefireopec

View topic pack schemas →

Cluster schema

Everything in one object

No stitching required — every field you need to act is present in the cluster response.

cluster_id — stable identifier across the full lifecycle
topic — the topic pack that owns this cluster
status — developing / breaking / fading / resolved
impact — 0–1 float aggregated from 7 sub-signals
headline — primary headline driving the cluster
affected_assets — exact tickers and FX pairs linked to this topic
why_it_matters — trader-readable summary (LLM-enhanced at impact ≥ 0.70)
ai_enhanced — boolean flag on every cluster
first_seen / updated_at — full timestamp trail for latency measurement

Start trading on signal, not noise

Connect to News Radar in minutes. One endpoint. Every breaking cluster. Scored, asset-linked, and AI-summarised.

Start building → Open dashboard

Loading QuantGist...

# Fetch breaking clusters with impact ≥ 0.65 from quantgist import QuantGistClient client = QuantGistClient(api_key="qg_live_...") clusters = client.news.radar( min_impact=0.65, status="breaking", ) # → [ClusterResult, ClusterResult, ...] print(clusters[0])

{ "cluster_id": "iran-war-20260528T0942Z", "topic": "iran-war", "status": "breaking", "impact": 0.85, "headline": "Iran Launches Drone Strikes on Israeli Military Bases", "affected_assets": ["XAU/USD", "WTI", "USD/JPY", "VIX"], "why_it_matters": "Escalation above threshold 3 — strait blockade risk...", "ai_enhanced": true, "first_seen": "2026-05-28T09:42:11Z", "updated_at": "2026-05-28T09:58:03Z" }