Vector search poor results

Questions
1

Hello community,

I am creating an AI assistant that answers user queries by retrieving information from a vector database. The issue I am having is that the search results are very low quality. I have another version using keyword search with HTTP tool that produces better results. I though vector search was supposed to be better, so I am sure I am missing something.

Share the output returned by the last node

Information on your n8n setup

  • n8n version:
  • Database (default: SQLite):
  • n8n EXECUTIONS_PROCESS setting (default: own, main):
  • Running n8n via (Docker, npm, n8n cloud, desktop app):
  • Operating system:
2

That usually has an influence from 3 different factors:

  • the model you are using to search
  • the quality of the vector store
  • the quality of the information in the vector store

Please try using a more powerful model like OpenAI 4o or 4o-mini and see if you get better results.

If not, then the issue is probably in the vector store. Review the settings you used to build it and the information inside.

Let us know how this goes! I’m curious to know what the solution will be!

3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.