Where to start
Don’t take our numbers. Go and get yours.
Connect a read-only token and we run one query over your own query history — about a minute. Or run that query yourself and paste the results back. Either way you get the same number: what repetition is costing the account, and which rules would capture it.
The quickest route. Paste a warehouse URL and a read-only token, and we run the query for you.
The scan runs one read-only SELECT over your query history. It writes nothing and reads none of your tables.
Your token is never stored. No database, no session, no cookie, and it is never written to a log. It is used for a single call to your warehouse and then it is gone.
It passes through a relay for one reason only: neither Snowflake nor Databricks sends CORS headers, so a browser cannot call them directly. Nothing about that hop retains anything.
The token is kept in your own browser so you don’t have to re-enter it, and Disconnect everything clears it along with any scan result. On a shared machine, use a short-lived token and disconnect when you are done.
Nothing connects to us on this route. You run the query in your own console and paste the results back. The arithmetic runs in this page, using the same shared analysis the connected scanner uses, so the two agree on identical rows.
-
1
Copy the query
One
SELECTwith aGROUP BYoverSNOWFLAKE.ACCOUNT_USAGE.QUERY_HISTORY, covering the last 30 days. Read it before you run it: it writes nothing and touches none of your tables.Snowflake SQL-- loading... -
2
Run it in your own console
Open Snowsight, pick a role that can read
ACCOUNT_USAGE, and run it. Then use Download results to export as CSV. -
3
Paste the results back
Column headers come with the export, so there is nothing to map. These rows never leave the page — open your network tab and watch.
Your export contains your real SQL text. If you would rather not paste that, drop the
QUERY_TEXTcolumn — you still get the full cost figure, but we cannot suggest rules without seeing the statements.