I have questions about using vector-based agents, and I’d like to ask for guidance on a specific issue regarding vector databases with agents.
My goal is to have my agent always access the Supabase vector database, but without it needing to be triggered as a tool or requiring the use of a description to call it. I want it to work similarly to the “Question and Answer Chain” node in N8N, where it automatically consults the vector database whenever needed.
Currently, I have a lot retriever nodes in the agent to allow users to query different types of content. However, I’d like to consolidate all this information into a single vector database that the agent always consults automatically, if relevant. I haven’t been able to figure out how to do this, but I believe it might be possible using Langchain code.
Does anyone have any suggestions on how to achieve this? Is it possible to configure the agent so it automatically consumes the vector database without being treated as a tool that requires explicit access instructions?
Information on your n8n setup
- n8n version: 1.68.0
- Running n8n via (Docker, npm, n8n cloud, desktop app): Docker