Duckpad

See the flow. Trust the result.

An open-source SQL workspace powered by DuckDB 2.0. Turn scattered data into visual pipelines, explore the results and trace a number all the way back to its source.

• Compose visually. Connect sources, joins and reusable components. Use guided forms to prepare columns and values, or write your own SQL nodes.
• Explore and check your data. Sample and filter results, save views, compute column statistics, and keep data tests and execution messages next to your SQL.
• Explain a number. Follow a result cell back to its contributing source values, with explicit lineage coverage.
• Review changes. Saves create Git commits. Work on separate branches, compare changes visually and review a merge before applying it.
• Run repeatable work. Share typed variables across nodes and environments, then launch Jobs against reviewed versions from the interface, CLI or a scheduler.
• Ask for help in context. Use the AI assistant to explain a flow, inspect its edits and confirm executions with your existing project permissions.
• Keep control. Run on your infrastructure, configure connections per environment, manage team access and import extensions for offline use.

Try it

With Python 3.11+ and Git installed:

pip install duckpad
duckpad start

Open the printed local address and sign in with the initial admin password from the terminal. Follow Duckpad in 15 minutes to build your first pipeline from scratch, write SQL and explain a result.

Prefer a ready-made project? Choose New project → Start from an example, enter a name, then Create → Run all to explore Northstar. See installation for platform requirements and setup options.

Go further

Build	Understand	Share and run
SQL syntax	Results and samples	Version control
Data preparation	Data tests	Jobs
Reusable components	Messages and logs	CLI and automation
Variables	Lineage	Connections
Project files	AI assistant	Environments

Installation & accounts · Single sign-on · Python nodes · Offline extensions · Project documentation · Dataiku & Alteryx vocabulary · Architecture · Validation · Releases · Security

Want to contribute? Start with CONTRIBUTING.md.