Log daily Mercury balances and alert when cash runs low

Every weekday at 8am, log my cash position across all of my Mercury accounts and warn me only when an account is running low. Use a cron trigger running Monday through Friday. Do not use a poll trigger: Mercury does not offer one, so a schedule plus a balance read inside the workflow is the correct pattern. Build this as a deterministic node workflow rather than an agent, because every step is known in advance and there is nothing to reason about.

Start with the Mercury List Accounts operation to read every account on the organization. Balances live directly on each account object as availableBalance and currentBalance, so there is no separate balance endpoint to call. Mercury returns monetary amounts as whole dollars in decimal form, where 12.50 means twelve dollars and fifty cents, so do not divide by one hundred and do not treat these values as cents.

Append one row per account to a Google Sheets history tab using the Append Values operation. Each row should capture the run date, the account name, the available balance, and the current balance. Append rows for every account on every run, whether or not any threshold was breached, because the whole point of the tab is a complete daily balance history I can chart later.

Expose a low balance threshold as a configuration input on the workflow, entered as a plain dollar figure, and apply it to every account. Compare each account's availableBalance against that threshold, not currentBalance. availableBalance is the spendable number and the two diverge whenever there is pending activity, so the low balance check has to use the available figure. An account breaches when its availableBalance is below the threshold.

If one or more accounts breach the threshold, send a single Slack message with the Send a Message operation to my chosen channel. List only the accounts that actually breached, and for each one give the account name, its current available balance, and how far below the threshold it is, so the shortfall is explicit and I know how much needs to move. Do not include healthy accounts in the message.

If every account is at or above the threshold, skip the Slack step entirely and end the run once the spreadsheet rows are appended. Do not send an all clear or a nothing to report message on those days. The alert is only useful if it stays silent when there is nothing wrong.

A Mercury Read Only token is sufficient for this workflow, since it only reads accounts and never moves money. That also avoids the IP allowlist requirement Mercury applies to Read and Write tokens, which makes setup considerably simpler.