v1 — public beta

Predict outcomes before they happen.

SignalGrid ingests public signal across news, social, video, search, and RSS, resolves it to entities you care about, and produces calibrated predictions you can act on.

Start free Watch demo

10verticals shipped

49RSS sources ingested

14prediction models in production

GET /v1/predictions/box-office/film_kj91

{
  "entityId": "film_kj91",
  "vertical": "box-office",
  "predictedValue": {
    "openingWeekendUSD": 78400000
  },
  "ci80": [68900000, 88100000],
  "ci95": [62100000, 96800000],
  "calibrationBucket": 0.78,
  "drivers": [
    { "feature": "trailer_view_z", "contribution": 0.14 },
    { "feature": "presale_velocity", "contribution": 0.11 },
    { "feature": "franchise_fatigue_index", "contribution": -0.06 },
    { "feature": "competitor_window_pressure", "contribution": -0.04 },
    { "feature": "marketing_spend_z", "contribution": 0.03 }
  ],
  "analogs": [
    "tt9876123",
    "tt8443209",
    "tt7012891",
    "tt6611123",
    "tt5500098"
  ],
  "asOf": "2026-05-09T14:00:00Z"
}

Every prediction ships with an 80 and 95 percent confidence interval, the five most predictive drivers, the five nearest historical analogs, and a calibration bucket telling you the model’s historical hit rate at this confidence level.

Trusted by teams making decisions on incomplete information

Ten verticals, one platform

One signal layer. Ten domains where predictions move money.

How it works

Ingest signals. Resolve entities. Predict outcomes.

Step 1

Ingest signals

Continuous ingestion across news RSS, social platforms, video, search trends, and public marketplaces. Deduped, enriched, and timestamped at the edge.

Step 2

Resolve entities

Every signal is grounded to a stable entity in our knowledge graph: a film, a candidate, a SKU, a creator. Cross-source coreference happens before scoring.

Step 3