Data Explorer

Every workflow now has a Data Explorer. Click Explore above the data table to spot-check and QA extracted data without exporting it first.

Two ways to query:

• Ask: type a question like "count of rows per brand, largest first". Kadoa writes the SQL, runs it, and shows the result.
• SQL: write DuckDB SQL against a single table called data. Column names autocomplete inline. Cmd+Enter runs.

Results replace the data grid and can be downloaded as CSV. The question box suggests examples built from the workflow's schema, such as counts per column and rows with empty values.

Queries run in your browser on the latest published run. Only the schema and five sample rows are sent to the model when you use Ask. Queries are read-only and show up to 1,000 rows.

Data Explorer documentation →