No usable API
QBXML through an SDK running against the file on Windows. Extraction is paged, rate-limited, and often has to run overnight against a restored copy.
By current system
QuickBooks Desktop has no modern API, the data lives in a file on a machine, and Intuit retires versions on a rolling schedule. That combination means Desktop users have less optionality than Online users and a deadline they did not choose — so the honest advice starts earlier and is more directive.
The situation
QBXML through an SDK running against the file on Windows. Extraction is paged, rate-limited, and often has to run overnight against a restored copy.
A .QBW on a specific machine, sometimes in a hosted Citrix environment, with a single-user lock that blocks work while anyone is in it.
Department, location, and often entity encoded into one field because Desktop offered nowhere else. Untangling that is the highest-value part of any move.
Intuit retires older versions on a rolling basis, and a discontinued version losing payroll and security updates is what turns a considered project into an urgent one.
Desktop values inventory with its own quirks. Reproducing the valuation exactly rather than approximately is a specific piece of work and a common source of unexplained COGS variance.
Customers, vendors, items, and jobs with duplicates, inactive records still referenced by transactions, and sub-customers used as a project workaround.
Less than for Online, and it is worth being straight about that. We can read a Desktop file on a scheduled basis — against a restored copy rather than the live file — and produce dimensional reporting, a shadow ledger, and AP capture on top of it.
What we cannot do well is real-time write-back. Online has an API that accepts finished bills; Desktop’s path for that is fragile enough that we would rather not build a workflow depending on it. So AP automation on Desktop means coding and approval happen with us and the bills are batched in, which works and is less elegant.
A Desktop conversion runs six to ten weeks and is quoted between roughly $28,000 and $55,000 depending on inventory, payroll history, and how many company files are being consolidated. That is roughly three times an Online conversion.
The reason to start two quarters early is not our capacity — it is that the work is decision-heavy. The class untangling and the chart of accounts design need your controller’s time, and those decisions cannot be compressed by paying more. A conversion begun under version-sunset pressure gets the technical part right and the structural decisions rushed, which is the outcome that produces regret.
In most Desktop files older than five years, the class list is encoding several dimensions at once. Splitting it properly gives you department, location, and entity reporting that Desktop could never produce. Carrying it across as-is means you have paid to move a workaround into a system that did not need one.
This is the part where an assessment earns its keep, and it is why we would rather look at the actual file before quoting anything.
Check it, and treat it as the real constraint rather than the aspiration. Losing payroll and security updates is what forces most Desktop moves, and the companies that handle it well started the diagnostic two quarters out. The ones that struggle started six weeks out and had to accept whatever chart of accounts design fitted the time available.
Questions
Version, entities, inventory, and payroll history is enough for us to give you a realistic timeline.