Unique's architecture is built around two core data flows, files and prompts, moving through a set of workloads that can include PostgreSQL for structured data and Qdrant for vector search. This page gives administrators a view of how these components fit together, as a starting point before diving into the RAG architecture, MCP Hub, and Conduct sections.

Data Flow

There are two main data flows in Unique: Files and Prompts.

The diagram below shows how these data types flow through the system. Depending on the setup, components such as PostgreSQL (the database) and Qdrant (the vector database) may be involved in the Unique Workloads.