libvector

Vector dot-product scoring — find semantically related content without a dedicated database.

Getting Started

import { calculateDotProduct } from '@forwardimpact/libvector';

Documentation

• Search Semantically — find related content by meaning with ranked results from a vector index (fit-rag search). You need no vector database.
• Give Agents Typed, Retrievable Knowledge — the full workflow to build an embedding pipeline from knowledge sources.

Internal CLIs

The fit-process vectors build step (in librag) turns resources into vector embeddings as one step that populates the index. It is not an agent-facing tool.