Knowledge
Documents, website pages and FAQs are split into passages, embedded and stored in a vector database. Agents with a knowledge base attached (Agent settings) look things up before answering.
| Knowledge base | Vector database | Embeddings | Documents | Passages | Agents |
|---|
Vector databases
| Name | Type | Target | Knowledge bases | |
|---|---|---|---|---|
| Built-in | Built-in (pgvector) | This app's Postgres, one table per knowledge base |