Where can I see what was spent?
Cell details: double-click a cell, or select it and press
Enter. Amount Spent shows what the latest run of that cell cost.Usage history: click History at the bottom right of the table, then Usage history. It lists one line for each charged cell. See Table options: row height, dependency graph, export, share and history.
Settings, then Usage: "Monitor your usage and spend." It has a Spend over time chart. See See where your spend goes on the Usage page.
Settings, then Plan & billing, then the Wallet tab: Funds usage & lookups. See Top up your wallet, set auto top-up and read your funds history.
The wallet button at the top right shows your balance. See Your wallet and keeping spend under control.
What starts a run that can spend?
A column spends from your wallet only when it is a paid one. In our tests, each of these started runs.
You start them yourself:
The column's play button. The Run column dialog opens first. On a paid tool such as Normalize Company Name, each of its buttons shows a Maximum cost line. That includes Run N rows with no result and Re-run N error rows, when the dialog offers them. On an AI column the line says Expected cost.
A row's play button (Run this row). It runs that row at once. There is no dialog and no price.
A cell's right-click menu. Run 1 cell starts that cell at once. With several cells selected, it says Run N cells. A paid cell run this way was charged.
Actions, then Enrich and override N rows. There is no dialog and no price. It re-ran every enrichment column on those rows, including cells that already had an answer, and paid cells were billed again. See Work with selected rows: enrich, archive or delete.
Re-run this cell, in a cell's details. An AI cell that was re-run this way was charged again, as a new line in Usage history.
Auto-run starts them:
A new value, with Auto-run on. A value typed into a column that a paid column reads started that paid column on that row. A value typed into a column it does not read started nothing. A blank row did not run until a value was typed. A column whose own Auto-run switch is off did not start.
Keep existing results, unticked. With Auto-run on, Keep existing results is ticked by default. Unticked, the app says: "Changing an input re-runs the columns below it and replaces their answers — which is billed again for those rows." Editing an input re-ran a paid column. Overwriting a column also re-ran a paid column that reads it.
A hidden column. A hidden paid column still ran on a new row.
Other things that charge:
Sandbox Mode. About sandbox mode says: "Steps run here are charged to your wallet exactly as they are live, on a few rows instead of every row." A paid row run in the sandbox showed Amount Spent. See Functions and Sandbox Mode.
Publish and run, in the Review changes window. The window shows no price. It ran the column that was added in that publish, and the run was charged. Publish and don't run ran nothing.
An AI web search that ends in "Unknown". A Web Search cell answered "Unknown" with Confidence "low". It was still charged.
PredictLeads Similar companies. When it runs on Loopster's account, the panel says: "Charged whenever the provider answers, including when the answer is negative." A row that showed "No lookalike companies found" was still charged.
What does not cost anything?
In our tests, none of these were charged.
Free tools and formulas. Formula, Normalize Domain and Wait added no lines to Usage history. A Formula column fills in on its own, even with Auto-run off.
Cells that show an error. A paid Normalize Company Name cell that said "Missing required input: Company Name" had no line. An HTTP (POST/GET) cell that said "API call failed: the request URL is not an allowed destination." had none either. The footer showed a count such as "1 cell failed".
Skipped cells. Cells that said "Run condition not met" had no line. See Run a column only when a condition is true.
Cells stopped before they ran. A paid cell that was still queued when you clicked Stop had no line.
Test prompt. The panel says it "Runs the Loopgent on the first row without saving results or spending from your wallet." Usage history and the wallet did not change.
A waterfall row that found nothing. A Work email waterfall row for a made-up person said "No email found", and its Amount Spent said "Free". Usage history listed "Reserved credits for managed <Provider> <action>" lines, each followed by a "Released managed-credit reservation for <Provider>" line with a negative amount, so each pair cancels out. The wallet button dipped by a few cents while a reservation was open, then went back. The app says a step that is missing a required input is "skipped (and not charged)". See Find people's contact details with waterfalls.
Editing a Use AI column's prompt. Saving a new prompt re-ran no cell and cost nothing, with Auto-run off or on, and with Keep existing results ticked or unticked. The cells kept their old answers, with no marker that they were out of date. Re-run the column to update them.
Troubleshooting
My wallet went down but Auto-run is off
What you see: Auto-run says Off, but the wallet button shows a lower balance.
Why it happens: Auto-run only decides whether new rows and values start runs on their own. The app says: "New rows and values will not run on their own — trigger them from a column or row. Cells already queued still run." A run you start by hand still happens. That includes a play button, Enrich and override, Re-run this cell, Publish and run and a run in Sandbox Mode.
What to do: Click History at the bottom right of the table, then Usage history. Read the times and Details of the newest lines, and match each one to something you ran. See Your wallet and keeping spend under control.
The Run column dialog price was higher than I was charged
What you see: The dialog showed Maximum cost or Expected cost. Usage history shows a smaller amount.
Why it happens: Maximum cost is a ceiling. With a run condition, the dialog priced all the rows, even though only some of them met the condition. The skipped cells said "Run condition not met" and were not charged. Expected cost on an AI column is an estimate. The amount charged can be lower than the estimate.
What to do: Open a cell and read Amount Spent, or open Usage history. They show what was charged. See Choose an AI model and see what it costs.
My spend jumped after I edited an input
What you see: You edited an input or clicked Overwrite all N rows. Usage history shows new lines for a paid column you did not run yourself.
Why it happens: With Auto-run on and Keep existing results unticked, the app says: "Changing an input re-runs the columns below it and replaces their answers — which is billed again for those rows." Overwriting a column re-ran a paid column that reads it and billed it again. With Keep existing results ticked, the app says: "Only errored, empty and new cells run. A completed answer is left alone even when one of its inputs changes." The paid column did not re-run then.
What to do: Click Auto-run at the top left of the toolbar and tick Keep existing results. It is also in Table settings, under Run settings. It only shows while Auto-run is on. Read the Run column dialog before you click Overwrite all N rows. With Auto-run on, it can say "+ ~$… across 1 downstream column it triggers". Free columns such as Formula follow along on their own and cost nothing.
I do not see a charge for a re-run in Usage history
What you see: You re-ran a cell. Amount Spent shows one small amount, and you cannot tell which line in Usage history is the re-run.
Why it happens: Amount Spent shows the latest run of that cell only, not a running total. Usage history has one line for each charged cell. A line has a Timestamp, Details, a Billing type and an Amount. It has no row number and no re-run marker. A re-run is a new line that looks like the first one. Free tools, errors, skipped cells and stopped cells have no line.
What to do: Match the times to when you clicked. The times are to the minute.
Good to know
Switching Auto-run on shows no price and no dialog. It ran nothing on the rows that were already in the table.
The app says cells already queued still run when Auto-run is off.
Before you run a paid column, check its price in the Add column panel and add it with Add column without running. See Run columns safely: Auto-run, test one row, then run the rest and Your wallet and keeping spend under control.