Running it
Where your books live
Three arrangements, and you choose.
On one computer. The database installs with the application. Nothing to configure, nothing to maintain, no server.
On your network. Several people in the office working against one database, on a machine you own. Point the application at it and it works.
On your own cloud account. For operations with more than one location. A database in an account in your name, anywhere that runs PostgreSQL. We never hold it and we are not in the path.
You can move between these as the operation grows. How you work does not change.
Users and permissions
Named users with roles. So the person entering bills is not the person who can void a posted check.
Permissions are enforced in the core of the application, not by hiding buttons. Hiding a screen is one thing. An operation a user is not allowed to perform is refused wherever the request comes from.
Access is granted per company. A user with rights to two of your four entities cannot see the other two.
Sessions can be listed and ended. The screen can be locked without closing the application.
Knowing who did what
Changes are recorded. Who entered a transaction. Who altered it. Who voided it.
So when a figure is not what somebody expected, "when did this change, and who changed it" has an answer instead of a theory.
Confidential records
Employee identifiers sit behind their own access control, with their own password.
A bookkeeper who needs to enter bills does not thereby get to read everybody's Social Security numbers. That access is granted on purpose, and separately from everything else.
Documents
Scans, photos of receipts, signed paperwork. All filed against the transaction they belong to.
Open the entry and the document is on it. So an auditor's question is a click, not a search through a shared drive.
Images are downscaled and re-encoded on the way in. A year of receipts stays a sensible size and copies quickly over a network.
The documents live in file storage, not inside the database. On this machine for a single user. On your own share on a network. In your own object storage online.
On a single machine, a backup is one archive holding both the books and the documents. It runs on a schedule and keeps the last several. So it does not depend on anyone remembering to run it.
On a network share or online storage, the documents are backed up with the rest of that infrastructure, using the tools you already run.
Barcode scanning works too, for entry where you would otherwise type identifiers off a label.
Backup and restore
Backups you take and hold.
Restore is tested as part of the same mechanism. A backup nobody has ever restored is a hope, not a backup.
Updates
The software checks for a new version and tells you. You install it when it suits you.
Never in the middle of a payment run because a vendor decided today was the day.
Learning it
Guided walkthroughs run inside the application, on the screen you are actually on. They step through a real task rather than describing one.
They make no changes to your data.