Hivemind custom trends allow users to define their own concepts and narratives and measure them consistently through time. By specifying keywords, phrases, or semantic descriptions, users can create bespoke trend signals that reflect portfolio-specific risks, themes, or investment theses.
Custom trends are point-in-time and constructed using the same underlying Hivemind ingestion and scoring framework as system-defined trends.
Custom trends can be created and explored in the ForecastOS UI, where users define concepts, review historical behavior, and compare custom narratives against existing Hivemind trends.

Custom trends are available through the ForecastOS API for automated creation, updating, and retrieval.
Detailed endpoint documentation is provided to clients with API access.
The open-source ForecastOS Python library provides helpers for submitting custom trend definitions and consuming the resulting time series in research workflows.
Concept embedding, scoring, and source processing are handled within the managed Hivemind platform.
Let's explore Hivemind Company Exposures next.