Skip to content
Geomacro

TECHNICAL PROOF · SECONDARY APPLICATION LAYER

Intelligence Pipeline

This page exposes implementation detail behind the broader ingestion and processing system. The public product presents three separate Risk Indices. Their current proof lineage comes from the versioned GRI v1.2 three-domain methodology: geopolitics, macro and rare-earth / critical-mineral risk. Other research or technical data streams must not be read as additional v1.2 scoring domains.

A working technical proof is not a claim of general production availability, production settlement, customer authorization or a service-level commitment.

Primary intelligence product
Data Pipeline

From source evidence to verifiable risk intelligence

The pipeline combines evidence intake, structured classification and event-level risk scoring. The current public GRI v1.2 calculation uses three scoring domains only: geopolitics, macro and rare-earth / critical-mineral risk. Other research and testnet streams remain separate.

Live technical view10 stages4 implementation layersGRI scoring · 3 domains
P1

Ingestion Layer

Source evidence to canonical event
  1. 01~5s
    Ingest
    Collect live source evidence across geopolitical, macro, critical-mineral and separate research streams.
    Multi-source evidence intake
  2. 02<50ms
    Normalize
    Reshape supported source payloads into one canonical event schema with provenance fields preserved.
    Canonical schema
  3. 03<10ms
    Dedupe
    Apply content and URL fingerprints to suppress obvious duplicate records before downstream scoring.
    Content + URL fingerprints
P2

Intelligence Layer

Relevance, classification and event risk
  1. 04<5ms
    Prefilter
    Apply relevance gates so unsupported or low-value material does not enter the accepted intelligence set.
    Relevance gate
  2. 05~1.2s
    Classify
    Route accepted evidence through the versioned classification layer for category, structure and provenance-aware metadata.
    Versioned classifier
  3. 06<200ms
    Score
    Assign event severity and confidence. The separate versioned GRI engine then computes eligible contribution deterministically.
    Event risk · 0–100
P3

Application & Feedback Layer

Optional forecasting and calibration research
  1. 07~800ms
    Predict
    Generate falsifiable forecasts only as a secondary application of accepted risk intelligence, separate from the GRI calculation.
    Secondary application
  2. 08rolling
    Reflect
    Evaluate prior calls and secondary market outcomes as calibration feedback without redefining the primary intelligence product.
    Calibration feedback
P4

Technical Proof Layer

Optional Arc attestation and testnet settlement
  1. 09~3s
    Attest
    Hash selected payloads and preserve an optional Arc Testnet attestation path as technical proof.
    Arc Testnet · SHA-256
  2. 10lifecycle dependent
    Resolve
    Within the secondary prediction-market application, resolve and settle testnet event contracts after the defined lifecycle completes.
    Secondary testnet layer