when you use Huggingface embeddings and

no, the model is running locally, on your machine

does it ever make sense to use a vector db for a single document?

In the context of information extraction

what would be the pros and cons of, for example, creating a llama index Chroma db set up to query against a 70 page .pdf vs. just sending that .pdf to Claude API and leveraging its large context window?

speed, runtime, citeability, cost

if none of those are important enough for the use case, by all means send the pdf to claude

consider that it would be a new .pdf each time

would that factor into it?

Youd have to measure the tradeoffs yourself I think 😅

for sure, I appreciate the feedback I'm just wondering if speed would actually increase if its a new pdf each time, same with runtime, citeability and cost

I cant say for sure without trying it myself

got it

ok thanks!