The rate card is not the price. A SERP at $0.0006 ($0.60 per 1,000) looks like pocket change until $50 has to sit in the balance before your first real search, your monthly block expires whether you used it or not, and the dashboard you bought still bills the same data underneath.
If you are budgeting with per-1,000 math only, add these three checks first. They change the real cash out without changing the per-request number.
The $50 That Sits Before Your First Search
Pay-as-you-go SERP providers typically show a pay-per-request price with no subscription. The same pricing pages list a separate line many buyers miss: a $50 minimum deposit once any trial credit is spent.
That is pre-funded balance rather than a fee (it pays for searches as you use them), but it is cash you park before you can test at production volume.
In practice it lands as one line: the workflow itself costs cents, then $50 has to be in the balance before you can run the first real one. The complaint is specific: a thousand results pages cost 60 cents in data, until you read the minimum top-up.
Two nuances matter:
- Trial without deposit. Some pay-go paths let you run a small free trial first (around $0.50 in test credits on wrapper-hosted plans) so you can validate the API before you park the $50 with the underlying data provider.
- Self-host does not remove the deposit. Going self-hosted removes the wrapper's $10 per month, but not the data provider's balance requirement. You still need that $50 sitting in the data account before the first search outside the trial.
For a sporadic workload (a few pulls here and there instead of a fixed $200+ per month) that $50 is not expensive, but it is illiquid. If you only needed 2,000 SERPs to test an idea ($1.20 at $0.60 per 1,000), you still front 40x that to start.
Expiry vs Rollover: Where the Money Actually Vanishes
Subscriptions and pay-go use opposite clocks. It is why a $129 block turns into a $449 upgrade when credits run out mid-cycle, and it is also where top-ups quietly win.
Subscription buckets expire. Fixed monthly plans give you an allowance: a starter around 200 credits near $29 per month, a solo-creator tier near $129, and comparable suite tiers near $139.95. What you do not use dies at the period end, so quiet months cost the same as busy months.
Top-up credits roll over. The wrapper's base subscription does not. On pay-go wrappers that bundle a dashboard over metered data, the structure splits:
- Base plan: one plan, around $10 per month. Includes about $10 of usage (a keyword search is around five cents, a domain overview eight, a brand check about $1.09, and ~217 weekly rank checks about $0.54). That included allowance resets each billing cycle.
- Extra usage: buy top-up credits as needed. Those top-ups roll over indefinitely and never expire on their own.
The FAQ line that catches buyers off guard is direct: while top-up credits roll over and don't expire, you need an active subscription to use the product at all. In other words, your banked data balance survives, but access to spend it requires the $10 per month to stay active. Unused base-plan credits do not roll over, only what you bought on top does.
Why this matters for budgeting:
- If your volume is seasonal or client-driven, a pay-go balance with rollover lets a light month stay light. A subscription bucket charges you for the ceiling you set in January, even if March is idle.
- If your volume is steady and you reliably burn the whole block, expiry is irrelevant; you paid for what you used anyway.
- A free tier of 250 recurring searches per month and one-time trials of ~2,500 requests can cover lightweight testing, but they do not change the expiry rule on paid blocks. The trial clock and the billing clock are different.
The Wrapper Doubt: Paying Twice for the Same Data?
MIT-licensed wrappers over pay-go data create a specific hesitation: paying for the wrapper feels like paying for the same thing twice.
Here is the split to evaluate without brand comparison:
- What the wrapper adds beyond the raw API: keyword research, backlinks, rank tracking, site audits, projects and memory, and agent ergonomics so you do not rebuild those workflows around the raw endpoint. The cleanest version of the trade-off: you lose the giant fixed subscription and can see exactly what each client is costing you. At ~$10 per month with $10 of usage included, the wrapper fee is visible per client, not pooled into an opaque suite seat.
- What it does not add: it does not create its own search index. The underlying data still comes from the pay-go provider and is billed per request. The MIT license is free; the data bill is not. The wrapper rents an index rather than owning a decade of crawl.
- The cost question at scale: the other worry is deduplication. Does repeated keyword fetching bill twice, erasing the savings when you run at scale? If you re-request the same terms without caching on your side, the meter runs again. Cache hits still burn quota on providers that count them, and nothing in a wrapper automatically deduplicates unless you add it.
When the wrapper wins: you want suite-like workflows without multiplying enterprise seats, you value per-client cost visibility, and the $10 per month saves you building memory, projects, and integrations yourself.
When direct pay-go wins: you only need raw SERPs for an app or agent, you already have your own storage and deduplication, and any dashboard is overhead you would not use.
Checklist: Uncover the Real Price Before You Buy
Use this against either model; subscriptions need it as much as pay-go does. The per-1,000 number only becomes comparable after these answers:
-
What must be parked before the meter starts? Confirm the minimum deposit after any trial, and whether self-hosting changes it. A $0.60 per 1,000 price with a $50 floor is $0.60 once you have parked the $50; it is not free to start.
-
What expires and what rolls over? Ask separately about the base allowance and about top-up credits. On subscription buckets, everything expires monthly. On pay-go wrappers, top-ups roll over indefinitely but the base subscription is still required to spend them; the included ~$10 resets.
-
Is there a free way to validate first? Look for a trial with a few cents of credit and for recurring free tiers (around 125–250 per month) or one-time trials near 2,500 requests. They let you confirm integration without committing to the floor.
-
What does the $29 or $129 actually cover? Starter blocks near $29 per month often cover only a couple hundred credits; the suite value is the UI and the historical data rather than the API calls. Heavy-use pressure that pushes a $129 account toward $449 is a credit-reset problem, not a per-request problem.
-
Where could the same request bill twice? If the provider counts cache hits against quota, repeated keywords cost twice. Plan your own caching and ask explicitly about deduplication rather than assuming the wrapper handles it.
qequals sits on the other side of this friction: pay-go shape, one $9 plan, nothing to park before the first call, and the expiry rule stated up front rather than in a FAQ (nothing carries over at period end, and a cached answer still counts). Hold it to the same checklist.
Ask what has to be parked, what expires, and where one request can bill twice. Then compare per-1,000.
