Data Engine

The Narrativ data engine has two primary components: nGress and Nexus.

nGress (Collection + Routing)

nGress is the collection and intake layer. It does three jobs:

Ingest from managed sources (APIs, scrapers)

Ingest from user-sourced sources (connected profiles, approved data classes)

Route inputs into the processing layer with traceable provenance (what came from where, under what permission state)

You can think of nGress as the system’s high-throughput intake valve that keeps inputs organized, attributable, and ready for computation.

Nexus is the processing layer where raw inputs become usable products:

Normalize raw inputs into canonical event and signal formats

Apply identity linking and de-duplication logic (where appropriate)

Compute features and aggregates

Produce indices with explicit definitions, time windows, and methodology versions

Publish outputs for B2B delivery, market resolution, and rewards accounting

Nexus is where Narrativ becomes market-grade: indices are computed in a way that is repeatable, versioned, and explainable.