Hivemind: Custom Pipelines

What Are Hivemind Custom Pipelines?

Hivemind custom pipelines expose the full set of tools used internally to construct company exposures.

They allow users to build end-to-end pipelines by combining structured data (from FeatureHub and elsewhere), unstructured data, and GenAI-driven transformations into custom, point-in-time derived signals.

Pipelines are fully configurable, with user-defined inputs, processing nodes / instructions, and outputs.

They can be used to create bespoke exposures, risk / alpha signals, or entirely new analytical artifacts beyond the standard Hivemind objects.

UI Access

Custom pipelines can be built and inspected in the ForecastOS UI, where users visually compose nodes, configure transformations, and access outputs.

API Access

Pipeline run creation and associated outputs are accessible via API.

Detailed endpoint documentation is provided to clients with API access.

Open-Source Access

The open-source ForecastOS Python library provides helpers for consuming pipeline outputs and integrating them into downstream research, risk, and portfolio workflows.

Next: Hivemind Risk Applications

Let's explore Hivemind Risk Applications next.