[perite]

This is a cool idea! A service that allows you to dump information and then query it with an LLM seems really useful!

It seems a bit expensive for the amount of storage you get though... Is there a reason for that? I don't know much about LLMs but do they need to generate a lot of additional data from the user's dataset for querying?

[k3ntaki]

The Ingestion process of RAG is very expensive, specifically when data volume increases, and requires lots of GPU processing