Skip to main content

Why Qdrant

Decision​

Use Qdrant as a vector database option for RAG and embedding-backed knowledge pipelines.

Context​

Agent systems need retrieval over project, client, and internal knowledge. A vector database provides semantic search, metadata filtering, and collection boundaries.

Consequences​

  • Collections should be named and governed by visibility and project scope.
  • Metadata must be designed before ingestion at scale.
  • Internal ops collections must remain separate from public-facing assistant indexes.