Platform / Market Intelligence

Explainable Intelligence

The same deterministic process behind every analysis — nothing invented, nothing hidden.

  1. 01

    Evidence

    Price and news evidence collected from real, attributed sources — fused and deduplicated before anything else happens.

  2. 02

    Reasoning

    Every item is classified as supporting, opposing, or unresolved — disagreement is surfaced, never smoothed over.

  3. 03

    Risk

    Assessed across eight distinct categories — market, event, liquidity, volatility, execution, evidence conflict, data quality, and uncertainty.

  4. 04

    Confidence

    Scored from how much real evidence exists and how well it agrees — never a number invented to sound authoritative.

  5. 05

    Explainable Analysis

    Presented in plain language — the evidence, the reasoning, the limitations, all visible, nothing hidden behind a single verdict.

Interactive Demo

See the shape of a real analysis

Click through each stage to see exactly what it produces — the same fields a live analysis returns.

Illustrative example — sample content, not a live market analysis

Supporting (illustrative)

  • Price action is holding above a recent support level.
  • Short-term momentum shows mild upward pressure.

Opposing (illustrative)

  • Volatility has increased compared to the prior session.

For technical evaluators — the exact seven services, in order

  1. 1

    lib/market-data/providers/alpha-vantage.provider.ts

    Market data ingestion

    Real-time price and news data pulled from external providers, normalized into a common evidence shape.

  2. 2

    services/ai/evidence-fusion.service.ts

    Evidence fusion

    Evidence from every source is deduplicated and merged before ranking ever begins.

  3. 3

    services/ai/evidence/evidence-ranking.service.ts

    Evidence ranking

    Remaining evidence is scored and ordered by relevance and source reliability.

  4. 4

    services/ai/reasoning/reasoning-engine.service.ts

    Reasoning

    Each item is classified as supporting, opposing, or unresolved — disagreement is surfaced, not smoothed over.

  5. 5

    services/ai/risk/risk-engine.service.ts

    Risk assessment

    Assessed across eight distinct categories, every time — never collapsed into one score.

  6. 6

    services/ai/confidence/confidence-engine.service.ts

    Confidence scoring

    Scored from how much real evidence exists and how well it agrees, not invented to sound authoritative.

  7. 7

    services/ai/explainable/explainable-analysis.service.ts

    Explainable output

    Everything above is composed into one transparent, human-readable analysis — nothing hidden behind a verdict.

Shares its evidence layer with the AI Assistant and Knowledge Base.

Open Market IntelligenceView Pricing