Narrative intelligence layer
See how narratives move, before they break.
Koryvant tracks how stories emerge and travel — emergence, propagation, trajectory — using hour-0-to-48 signals, so newsrooms, brand teams, and intelligence desks see the next move, not just the last one.
Method
A falsifiable process, not a dashboard claim.
Koryvant treats narrative forecasting as a scientific question: freeze rules first, then test under prospective conditions with transparent evidence constraints.
Freeze Before Intake
Methodology, schema, and outcome rules are declared and locked before candidate selection starts.
Mechanical Selection
The first qualifying candidates after a committed timestamp are drawn in order, with no discretionary filtering.
Sealed Prediction Ledger
Each prospective prediction is committed with packet hash and timestamp before outcomes exist.
Pilot
Forecast Pilot v1 execution path.
The operating target is five clean prospective registrations with zero protocol violations. Accuracy is measured after procedural integrity is proven.
Anchor Commit
Publish seed, intake start timestamp, and frozen artifact hashes before any narrative is selected.
Intake Log
Record every qualifying narrative crossing the velocity floor with UTC timestamp and source provenance.
48-Hour Packet
Build packet, assign birth type and T0, and generate four window predictions with confidence values.
Resolution
Fill outcome blocks in later commits using the pre-locked winner, failure, and ambiguous criteria.
Evidence
Current synthesis baseline.
Five narrative case files currently form the reference set. The discriminators are promising, but sample size remains intentionally constrained until further falsification rounds.
Institutional
Altman / OpenAI board event
Winner trajectory with persistent velocity and multi-cluster amplification within the first 48 hours.
Hybrid
SVB collapse
Reflexive narrative where social propagation affected real-world outcomes and mainstream spread dynamics.
Social
Bud Light backlash
Cross-cluster bridging and durability effects highlighted long-tail propagation behavior.
Social
Vision Pro returns + Greptile workweek
Failure-class narratives helped isolate counter-strength, concentration, and decay as key separators.
Research updates, beta requests, and collaboration
For pilot status, beta access requests, research discussion, or institutional collaboration, write to hello@koryvant.com.