Have your agent build a shared chart
Give your agent a task: inspect your Stripe data, calculate an invoice count, save the query, create a chart, and publish it. This walkthrough builds a chart of paid invoices by creation month from Stripe. It follows the example on the homepage. The count includes invoices marked paid, including zero-value invoices; it does not measure revenue or MRR.
You need a rawquery account, access to a Stripe account with invoices, and permission to create a read-only API key. For the agent route, use Claude in your browser, Claude Code in your terminal, or Codex in your terminal. Their accounts and usage are separate from rawquery.
1. Create your workspace
Start free, then sign in. No credit card is required. Open Connections in your dashboard and select Add Connection.
2. Connect Stripe
Choose Stripe, enter your restricted read-only key, and name the connection my_stripe. Select the invoices stream and complete the connection setup. Keep the key in the connection form rather than pasting it into a chat.
Wait for the first successful sync in Connections. Its duration depends on the data volume and Stripe's API limits. Open Lakehouse and check that my_stripe.invoices contains rows.
The connection name becomes the schema name in queries. If you used another name, substitute it below. Subsequent syncs follow the schedule you configure.
3. Ask for your chart
Choose where your agent works:
- Claude in your browser: connect your workspace, then enable rawquery in a conversation.
- Claude Code in your terminal: install rq and the rawquery skill, then start Claude Code in that project.
- Codex in your terminal: use the Codex setup in the same guide to install its rawquery skill, then start Codex in that project.
Give your agent the complete task:
Inspect the my_stripe.invoices table. Count invoices where paid is true,grouped by their creation month, ordered oldest first.Show me the SQL and the result. Save the query as paid-invoices-by-monthand create a bar chart: month on the x-axis, invoice_count on the y-axis.Name the chart paid-invoices-by-month and title it Paid invoices bycreation month. Format month as YYYY-MM. Do not publish it yet.Inspect the my_stripe.invoices table. Count invoices where paid is true,grouped by their creation month, ordered oldest first.Show me the SQL and the result. Save the query as paid-invoices-by-monthand create a bar chart: month on the x-axis, invoice_count on the y-axis.Name the chart paid-invoices-by-month and title it Paid invoices bycreation month. Format month as YYYY-MM. Do not publish it yet.Your agent uses rawquery's tools to inspect the table, run SQL, and create the saved query and chart. Review its tool requests and output. The chart should appear under Charts in your rawquery dashboard.
Using the SQL editor instead
Open Query and run:
SELECT strftime(created, '%Y-%m') AS month, COUNT(*) AS invoice_countFROM my_stripe.invoicesWHERE paid = trueGROUP BY 1ORDER BY 1SELECT strftime(created, '%Y-%m') AS month, COUNT(*) AS invoice_countFROM my_stripe.invoicesWHERE paid = trueGROUP BY 1ORDER BY 1Save the query as paid-invoices-by-month. Open Charts → New Chart, select that saved query, choose a bar chart, and set month as X and invoice_count as Y. Save the chart.
The query uses the invoice creation date, not the date payment arrived. If it returns no rows, check the sync status and whether your Stripe account has invoices marked paid.
4. Review and share
Check the dates and counts in the chart. To share through your agent, ask:
Publish the paid-invoices-by-month chart with password protection.Ask me for the password before publishing.Publish the paid-invoices-by-month chart with password protection.Ask me for the password before publishing.Or publish from the chart's actions in the dashboard. Publishing creates a public URL. Anyone with the link can view an unprotected chart; use a password when access should be limited. You can unpublish it later.
A chart reads synced data. Configure source syncs and the chart's refresh interval if you want a recurring report. See chart refresh and access settings.
Open the existing published example. Its source connection is named stripe_live, so its counts reflect that workspace rather than your account.
Next
- Match HubSpot contacts to Stripe customers.
- Combine charts on a dashboard.
- Connect another source.
- Use Claude Code or Codex, or install the CLI, if you prefer working in a terminal.
For help with a connection or query, contact hello@rawquery.dev.