Vector store
A vector store saves mathematical representations of information. It is used with embeddings and retrieval so AI can answer using your documents.
On this page
Source content checked . English translation published 2026-10-27.
In business automation
A vector store lets you save representations and search for similar items. It is sometimes used in the retrieval stage of RAG, alongside information that identifies the document, its permissions, and when it was updated.
Examples
These examples illustrate a possible workflow. Adapt the connections and rules to your own systems.
Business process example
Company procedures are divided into sections and saved with a document ID and department permissions. For a new question, only sections the employee is authorized to read are retrieved, and an answer is then prepared with a reference to the procedure.
Relevant tools
- Supabase
- Can combine data storage and vector search; you need to build the ingestion and permissions filtering process.
What to check
- Do not store only the vector. Without metadata, it will be difficult to remove an outdated document or filter information by permissions.