ExplainabilityLive

The explanation is the reasoning

Not a summary bolted on afterwards. Every recommendation is a chain of signed, weighted, provenance-tagged evidence you can follow end to end.

ExamplePOST /v1/reason
Recommendation · AAPL
+0.62bullish
net_pressure · confidence 0.71
Evidence paths
inflation ↑
+0.42
real_rates ↑
-0.28
usd_liquidity ↑
+0.31
credit_spread ↑
-0.14
Example output — sign in to run live reasoning on your own scenarios.
Evidence paths

Follow any number to its source

  • Each contribution carries a sign, a magnitude and a provenance record.
  • Inferred edges are marked distinctly from observed ones.
  • The same paths power /v1/evidence — the UI shows exactly what the API returns.
ExamplePOST /v1/reason
Recommendation · AAPL
+0.62bullish
net_pressure · confidence 0.71
Evidence paths
inflation ↑
+0.42
real_rates ↑
-0.28
usd_liquidity ↑
+0.31
credit_spread ↑
-0.14
Example output — sign in to run live reasoning on your own scenarios.

Why it matters

Auditable

Defend a decision by showing the evidence behind it.

Debuggable

See which factor moved a verdict, and by how much.

Honest

Confidence is surfaced and labelled as uncalibrated.

See the reasoning for yourself

Create an account and explore explainable recommendations in the interactive workspace.