Leonardo V.

Hey Nick C. ! We’ll file this with our UI team - will let you know here for updates

Thank you Georg H.!

Hey Georg H. - what are you currently building on MotherDuck that would require a business plan without the platform fee?

This is great Nick C.! We will be adding to our docs as well!

DMing you to talk more about this

Hey Nick C., you are correct and option 1 does not work. We do not (yet) support async/long-running remote tasks without an active client. If you need to avoid the 55s timeout limit of our remote MCP, my suggestion today is to spin up a local DuckDB MCP server using: https://github.com/motherduckdb/mcp-server-motherduck . Though it requires local CC and I guess this does not work for you?

Hey Nick C.! Also looking into this with our AI team. If we would publish the extension via an accessible method (e.g. pypi), and it'd be downloaded in the sandbox, would the sandbox be able to connect and talk to MotherDuck?

That is correct Tim M. - so you'll have to keep in mind that if underlying Iceberg files get modified but new operations these changes will not be picked up by the DuckLake. This approach is great if you update your data in batches on an interval. In that case, you could very easily append a COPY operation after each pipeline to "refresh" the DuckLake database

I think parquet would fit better this use case!

Hey Zak! Agree with the above: this is a sensible approach if your platform team does not want for pipelines that touch production systems to have wide internet access. You then would have: • One pipeline from prod systems -> Azure blob • One pipeline from Azure blob -> MotherDuck