MOTHERDUCK
DuckDB in the cloud.
AI in the loop.
MotherDuck brings DuckDB to the cloud. camelAI writes the SQL, builds the dashboards, and publishes the results — so your team can stop waiting and start analyzing.
Already know DuckDB? Meet its cloud layer.
DuckDB, but in the cloud
Your favorite analytical database, now with cloud storage and sharing. Same SQL dialect, same speed — accessible from anywhere, by anyone on your team.
Hybrid execution
Queries run locally AND in the cloud, automatically choosing the fastest path. Small dataset? Runs on your machine. Petabyte scan? Offloads to the cloud. You don’t think about it.
Zero infrastructure
No clusters to manage, no nodes to size, no auto-scaling policies to tune. Connect and query. That’s it.
The numbers speak for themselves.
And camelAI writes the SQL for you.
Native DuckDB SQL. Not generic SQL with a DuckDB adapter.
camelAI writes idiomatic DuckDB — read_parquet, QUALIFY, GROUP BY ALL, QUANTILE_CONT — because it understands the dialect, not just the standard.
S3 Parquet Analysis — Product Revenue
Analyzing revenue by product line directly from Parquet files in S3 via MotherDuck. No import step — read_parquet queries the files in place.
SELECT product_line, SUM(revenue) AS total_revenue, COUNT(*) AS orders, QUANTILE_CONT(revenue, 0.5) AS median_order FROM read_parquet('s3://warehouse/sales/*.parquet') WHERE order_date >= '2025-01-01' GROUP BY ALL QUALIFY total_revenue > 100000 ORDER BY total_revenue DESC;
Finding: Enterprise accounts drive 47% of total revenue despite being only 12% of orders. Median order value is 8.2x higher than Growth tier. Recommend prioritizing Enterprise sales motion for Q3.
What will you ask your MotherDuck data?
“Connect to our MotherDuck database and build a dashboard from the sales schema. Show revenue by region, top products by margin, and a 12-month trend line. Publish it for the team.”
Try this prompt“I have a set of Parquet files in S3. Query them directly through MotherDuck — no import needed. Show me summary statistics, distribution plots, and flag any data quality issues.”
Try this prompt“Run a cohort retention analysis on our user_events table. Use DuckDB’s window functions to calculate weekly retention curves by signup source. Build a publishable report.”
Try this prompt“Set up a daily cron job that queries our MotherDuck database for new anomalies in the metrics table, generates a summary notebook, and emails it to the data team.”
Try this promptFast data deserves fast answers.
Start analyzing your MotherDuck data today.