Adithya K.

In reality, there shouldn't be any restrictions on using MD there. Two things that might interfere in getting started would be the pinned DuckDB version in that LangChain extension and possibly getting the right database to be queried.

Hello Sugendran, LangChain supports DuckDB at the moment, and not yet MotherDuck. To create a knowledge base in LangChain you can use this: https://api.python.langchain.com/en/latest/vectorstores/langchain_community.vectorstores.duckdb.DuckDB.html and here is a blog on building a knowledge base on LlamaIndex https://motherduck.com/blog/search-using-duckdb-part-2/

Hey Mark, from what I understand from your question is that, for capturing changes in the database and updating it into MotherDuck we recommend Estuary (docs here) for our users. Most of our integrations are mentioned here: https://motherduck.com/docs/integrations/ .

Hey Rob, also this seems like a great question for the DuckDB community folks too. You can find their discord channel here: https://discord.com/invite/tcvwpjfnZx

Hey Mark, thanks for the question. I'll ask the team if they have an answer for this.

Hey Rob, that is an interesting question. To understand your question better, would it be to only read files, or also analyze/process them? The limitation would be that DuckDB is focussed on local processing.

Hey Kirk, I suppose your question relates to hosting your S3 files. In terms of performance, I'd say it does matter for the files to be in the same region, and we strongly recommend it be in us-east-1. Especially so when the files are large.

Hello John, those are some very interesting connection tests you've done there. Thanks! I'll convey it back to the team. On the token front, if I understand you right, you are referring to token based access control, where each token has rules on objects, right?