DB Pilot

The AI assistant sits inside the SQL editor, the table viewer, the Python blocks, and even the chart builder. You can ask it to generate a query, explain one you didn't write, or fix an error, and it uses context from the databases and tables you choose. It is a natural language to SQL shortcut for people who know roughly what they want but not the exact syntax. It supports the latest models from OpenAI, Anthropic, Google, and xAI, plus custom models from any provider once you add your own key.

Notebooks are the part that sets DB Pilot apart from a plain SQL client. Each block holds SQL, Python, text, or a chart, and they run in order like a Jupyter notebook. Query results are available in later Python blocks, and Pandas or Polars dataframes are available in later SQL blocks, so you can move between the two languages without copy and paste. You can parameterize blocks with Jinja and store secrets as environment variables.

You can turn a notebook into a small dashboard with input fields, date filters, tables, and charts. That means the same file serves as both your scratch pad and the report you hand to someone else. It's lighter than a BI tool on purpose, aimed at a repeatable analysis rather than a company wide dashboard.

The table viewer lets you browse records and change them the way you would in a spreadsheet. You can insert, update, and delete rows, and filter or sort tables without writing SQL. For anyone who owns a small database, that removes the fear of a mistyped DELETE.

The editor offers context aware autocomplete for tables, columns, keywords, and functions, so you spend less time recalling exact names. Paired with the assistant, the pane works as an AI SQL editor: describe the result you want, and the query appears for review. It also formats your SQL to keep it consistent, keeps a query history you can re-run, and lets you save favorites for the statements you run often.

All app data and connection credentials stay on your machine, with passwords stored in the system keychain. Database connections happen directly between your computer and the server. The AI assistant can see which tables and columns exist so it can help, but it doesn't read the actual rows in your database.

You can export results or copy them to the clipboard as CSV, JSON, NDJSON, or Markdown, which covers most handoffs to colleagues. The chart builder lets you build visuals without code and compare values across time periods from inside the results grid.