Why we show the cost before you click, not after
Most credit systems tell you what something cost after you've already paid for it. Ours tells you first, holds the credits before the vendor is even called, and gives back the difference if the real cost came in lower.
There's a specific moment a lot of SEO tools have in common: you run a search, close the tab, come back later, and the number in the corner is smaller than you remember. Something happened between the click and now, and you weren't told about it when it counted.
That's not a bug in those tools. It's the shape of a credit system that checks the cost after the work is done instead of before.
The failure we built against
The pattern shows up in compiled sentiment about the big suite tools often enough that it's not an edge case: people burning through a month of credits in a single research session with no warning, then locked out of their own work until the balance resets — "particularly frustrating," as one summary put it, "since they've already paid." You didn't overspend on purpose. You just couldn't see the meter while you were using it.
A cheaper credit-metered tool doesn't fix this. It's the same shape of anxiety at a lower price — you still don't know what a session is going to cost until it's already cost it.
What happens instead, on every paid button
Every action that spends a credit on Nextrank runs the same four steps, in this order:
- Estimate. A calculation from the parameters you're about to submit — no network call, no charge, and it's on the button before you click it.
- Reserve. If you click, the credits are held on your balance before we call the data provider. Can't afford it? The action is refused right here. You are never charged for a call that already went out.
- Fetch. We call the vendor.
- Settle. The hold becomes a real charge for what the call actually cost. Came in under the estimate? The difference goes back. Call failed? The whole hold releases.
Two specific things become impossible once you build it this way: getting billed for something that didn't work, and finding out you overspent after the money's already gone. That's it. That's the whole mechanic — see the full breakdown here if you want the exact cost table.
The other half: plans and credits never mix
The reserve-then-settle sequence only works because it's metering one specific thing — discretionary research (keyword search, metric refresh, ad-hoc lookups). Scheduled rank tracking runs on keyword slots that come with your plan, and touches credits exactly zero times. A zero credit balance never stops a tracker that's already running. Two currencies, one job each, never blended — that split is deliberate, not an accident of how the billing happened to get built.
What this costs, in practice
100 credits = $1. A 250-row keyword search is 21 credits — you see "21 credits" on the button before you touch it, not a debit you notice afterward. There's no plan-tier tax on top of that number, because the number already is the whole story.
Compare that to what happens when the meter is hidden until the invoice: you either over-provision "just in case," or you get cut off mid-project and find out your work is paused until renewal. Neither of those is a pricing problem. They're a visibility problem, and visibility is the one thing a credit system can actually promise.
If you're currently paying suite prices for a workflow you use narrowly, here's what that actually costs broken down against a per-action model. And if you want to see the estimate mechanic yourself before deciding anything, the pricing calculator runs the same math the product does.