Trust center
Signals first. Claims second.
Dispel is built around evidence inspection, calibrated uncertainty, and practical next-step guidance. It does not market mind reading, biometrics, or certainty where the evidence is weak.
Evidence model
Dispel scores provenance, scam pressure, manipulation indicators, and claim consistency using the evidence actually inspected.
Unknown states
When providers are absent or evidence is weak, Dispel should say Unknown instead of inventing confidence.
Workspace security
The product supports password-backed workspace authentication and signed sessions, with operator-facing diagnostics for deployment readiness.
Provider handling
Remote providers can be enabled per medium, but provider output is normalized into the Dispel score model rather than passed through as raw claims.