Has anyone used LlamaIndex?

Has anyone tried LlamaIndex? Looks like it is quite powerful albeit the user interface may be a bit crude. Any user stories?

1 Like

LlamaIndex is great. If anything it’s some great examples on how to parse and collect information

I’ve followed this (and a couple of other similar) tutorials and built a chatbot based on our own website pages as source, and it works pretty well, but I have no idea how to customise the input data! It’s a bit slow to respond to a query too, but I don’t know if that is to do with running it locally or not? Not production speed for sure…

LlamaIndex works great especially for proof-of-concept demos. It internally uses LangChain, which is also great. We used both libraries before starting to build Gista with my own implementation. I learned a lot from those libraries previously.

My recommendation is to start with LlamaIndex because it’s easier, and as you gain experience in internals, dive into LangChain, which is a giant set of utility functions where you pick your subset for your use case.

I’m working with Llamaindex. Data in Memory works fine.

Does Llamaindex slow down performance with large volumes of data

I’m trying to use it with an input text document containing 5000 rows of data and it works great on the first 20 rows or so of data. IOW, my queries don’t return results for rows past 20. Maybe I misunderstand how to use LlamIndex on large datasets.