Build your first dashboard

Build a small dashboard with a SQL query, a chart, a filter, and a URL you can share. You describe the outcome; your agent uses BitBoard’s MCP tools to create it.

This assumes you’ve already connected an agent. If you haven’t, start there.

Create a dashboard

Tell your agent:

Create a new BitBoard dashboard called “first run”.

The agent creates the dashboard and returns a URL. Open it and keep both windows visible so you can watch changes appear.

Import a source

Use this public global GDP dataset:

https://raw.githubusercontent.com/datasets/gdp/master/data/gdp.csv

Tell your agent:

Fetch that CSV and import it into the dashboard as gdp.

dashboard_import_csv adds the data to the dashboard catalog and creates a starter SQL block. The agent should inspect that source before building the final view.

Build the dashboard

Tell your agent:

Inspect the gdp source and build the dashboard in one update. Reuse or replace the starter query so there is only one final SQL block. Add a country filter, a yearly GDP line chart, and a short heading. Arrange the heading and filter above the chart, then run everything and verify the result.

For a multi-block change, the recommended path is one declarative dashboard_update containing the SQL, chart, filter, displayed state, and layout. This avoids a sequence of one-block and one-layout calls.

The agent should inspect the catalog, confirm the columns, apply the update, run the graph, and check the rendered result. You do not need to memorize the tool sequence, but you can inspect the SQL and block sources in BitBoard.

Verify it

Open the dashboard and confirm:

• the heading, filter, and chart are visible;
• changing the country updates the chart;
• the SQL block reads from the imported source; and
• the chart reads from the SQL output.

If something is wrong, describe the visible result to your agent and ask it to inspect the relevant block before updating it.

Share the dashboard

When the result is ready, tell your agent:

Make the dashboard viewable by link.

The agent sets visibility to link_view. Anyone with the URL can view the dashboard anonymously and read-only; editing still requires authentication and permission.

Where to go next

To use a live database or business app, see Connect your data and apps.

For the full block reference, see Blocks.

For the dashboard tool reference, see Dashboard MCP tools.