Choosing a CSV generator when six clients share one desk

Three decisions sit in front of you before you spend anything: which lifetime tier you buy, how many of your six or seven clients it should serve, and who reads each file before a client sees it. The budget question is really a scope question, because the tiers are licences for the same listing work rather than separate products.

For an agency, the useful measure is how many client batches you can produce per generation run and still cover with the credit allowance attached to your chosen tier. Bulk AI listing generation does the row drafting, and the major store exports give each client a file in a format their platform accepts. That is the work that would otherwise sit on a junior's evening.

Keep separate client catalogues as separate runs. The objection you raised at the start is a shared inbox mixing clients together, and the answer is process rather than promise: name every export for its client, keep source sheets apart, and never let one client's generated rows sit in a folder another client can see. Your budget then buys drafts plus your review time.

Three lifetime options are published on softwareprogrammer.app: Creator, Store and BYOK Store. Each is a one-off purchase, so read them as different amounts of the same listing capability rather than a monthly plan to weigh against a subscription. Exact figures are listed in the publisher's own price table, not here.

Creator suits an agency testing the output on one or two clients. Store widens what you can do for clients whose catalogues need the store exports and SERP title variations as part of routine delivery. BYOK Store covers agencies that bring their own key to the generation step, which changes how a heavy catalogue workload is handled.

Before you pick, match the tier to your real pattern across six or seven clients. If two clients send monthly batches and the rest send one catalogue a quarter, size for that combined load rather than for a busy month of trial. The tier that fits is the one whose credit allowance and export options cover your ordinary month without stretching.

The doubt is not really about price. It is about keeping client sending separate, and that comes down to how your team names, stores and reviews files, not how the tool behaves. ShopCSV AI produces rows and exports; your desk decides which client each batch belongs to and who signs it off.

You can start small. Trial credits let you put a real client catalogue through the generator before any lifetime tier is bought. Take one client whose catalogue you know well, generate a batch, export it, and check the result against that client's own store rules. If the rows need too much correction, you have learned that on the trial.

The lifetime commitment is not risky in itself, because it is one payment rather than a renewal. The risk sits in buying a tier above your actual workload. Test the trial credits first, then buy the tier that matches the six or seven clients you genuinely run, and keep the review pass on your side of the line.

Decide in one order: trial first, one client catalogue second, tier choice last. Run trial credits on a batch you already have a known-good sheet for. When the generated rows and the store export come back, compare them with what your team would have produced by hand. That comparison, not a feature list, tells you whether the tool saves your reviewers time.

If the trial batch needs heavy editing, buy the smaller tier or wait. If it clears review with light edits and the export opens cleanly in the client's platform, move up to the tier that covers your whole client list. Record which tier you bought and for how many clients, so the next renewal conversation has a reference point.

The upgrade decision is also a workflow decision. Add clients to the generator only as fast as you can keep their files separate: one folder, one naming rule, one reviewer per client. Spreadsheet exports do not carry client boundaries, so your process does. Settle that process before you buy the larger tier, and the shared-inbox objection stops applying.