Thanks for sharing, evidence looks very interesting. Reading the docs, do I understand it right that I need to publish parquet files so that WASM DuckDB could query them. I’m a bit concerned users wholesale downloading all the data.