Credits and metering
RankX AI meters every action from one credit wallet per account. Credits are reserved first, consumed on success and released in full if the work fails.
RankX AI meters work with credits, from one wallet per account. Every action that calls an outside provider follows the same lifecycle: credits are reserved before the work starts, consumed if it succeeds, and released in full if it fails. You are never charged for work that did not produce a result.
There are no per-feature quotas. There is one balance, and you decide what to spend it on.
How does a charge work?
Every charge is reserved first and settled only when the work is done. If the balance is short, the action is refused before anything is bought, with the reason.
graph LR
A[Action requested] --> B{Enough credits?}
B -->|No| R[Refused, with the reason]
B -->|Yes| C[Reserve]
C --> D[Provider call]
D -->|Success| E[Consume the reservation]
D -->|Failure| F[Release in full]Two properties follow from that shape:
A failure costs nothing. If an assistant times out, a crawl cannot reach your site or an article run delivers no draft, the reservation is released and your balance returns to where it was. An article whose draft was saved before the run stopped is yours to edit, and its charge covers the writing steps that completed.
A partial run is honest about what it did. If a scheduled run is interrupted or the balance runs out halfway, RankX AI charges for the checks it completed, marks the run partial and reports the shortfall. It never returns a half-finished result as a whole one.
What is charged, and per what?
The unit matters more than the price. Two examples cover most of the bill:
An AI visibility check is charged per prompt, per assistant, per run. Twenty-five prompts on ChatGPT alone is 25 charged checks on every run. With three assistants switched on it is 75, and with four it is 100. Multiply by your cadence to get a monthly figure.
An AI Shopping check is charged per shopping prompt, per surface, per run. Shopping prompts are a separate allowance from AI visibility prompts, and the two do not draw on each other.
Every action, with its unit and current price, is in the credit cost reference, which is generated from the product. Prices are not repeated anywhere else in these docs, because RankX AI publishes rates in versions and a number typed into prose goes out of date with the next version.
What does not cost credits?
Reading is always free. Every report, dashboard, export and read tool on the MCP server costs nothing to use.
A cache hit is free. RankX AI keeps keyword research results for 14 days and keyword trends for 7 days, and repeating a recent request inside that window returns the stored result without a charge.
Setup is free. The research and scanning in the setup wizard are funded by RankX AI and do not touch your balance. See the free trial.
How are plan limits different from credits?
Plan limits cap how much you can set up; credits cap how much work gets done. The two behave differently and have different remedies.
| The wallet | Structural limits | |
|---|---|---|
| What it is | A balance of credits | Per-plan ceilings on setup |
| Examples | Every check, crawl, brief and draft | Websites, tracked keywords, prompts per Website, seats, Client Workspaces |
| What it caps | How much you can spend | How much you can set up |
| When you hit it | The action is refused with a credit reason | The addition is refused with a limit reason |
| How to raise it | Add credits | Change plan |
Structural limits are enforced in the database, so a bulk import that goes over one is refused rather than silently trimmed. The figures are in plans and limits.
Because one balance covers every metered action, a per-feature quota such as "20 manual checks a month" does not describe how RankX AI meters.
Why are prices versioned?
RankX AI keeps its prices in a versioned rate catalogue. A published version never changes: a new price is published as a new version, so the price behind every past charge stays knowable.
That is why prices are resolved live everywhere they are shown, including in the descriptions of the MCP tools that spend, and why these docs point at one generated page instead of repeating figures.
Why was a spend refused?
RankX AI gives one of four refusal reasons, and each has its own remedy. Keeping them apart means you are never told to buy credits when credits are not the problem.
| Refusal | What has happened | The remedy |
|---|---|---|
| Not enough credits | The wallet is short for this action | Add credits, or reduce recurring spend |
| Subscription not active | The subscription has lapsed or been cancelled | Reactivate it in billing |
| Client budget exceeded | An agency client has spent the ceiling set for it | The agency raises that client's allocation |
| Action not priced | RankX AI has no price for this action | Nothing to change on your side. The message says so |
The third applies only to agencies, and the agency wallet can be full while a client is stopped. A per-client budget is a ceiling on the shared wallet, not a separate pot, so topping up the agency wallet changes nothing for that client until the allocation is raised. See per-client credit budgets.
How do I control what I spend?
Recurring spend in RankX AI comes almost entirely from four things, in this order:
- Active prompts, multiplied by enabled assistants, multiplied by how often they run. This is usually most of the bill.
- Tracked keywords, multiplied by the check cadence.
- Active shopping prompts, multiplied by two surfaces and the cadence.
- Everything else, which is occasional rather than recurring: audits, research runs, briefs and drafts.
The levers that work:
- Pause prompts you do not act on. A paused prompt stops costing credits immediately, keeps its history and can be switched back on at any time.
- Match the cadence to the plan. Each plan's monthly grant is sized for a cadence that answers real questions. Running daily when weekly answers the question is the most common cause of an unexpected balance. See check cadence.
- Untrack keywords you do not act on. The keyword's history is kept, and the slot is freed against your plan limit as well as the spend.
- Check the cost before a big run. Every screen that spends shows what an action will cost before you start it.
Where do my credits go?
The usage report groups spend by action or by Website over a period, so a drop in your balance always has an explanation. Work that failed and was refunded is left out, because it was never spend.
Each monthly grant covers its own billing period and does not roll over, which is another reason to match cadence to plan rather than buy headroom you will not use.
Where to go next
- Credit costs, the generated per-action table.
- Check cadence, for how often checks run.
- Plans and limits, for every structural ceiling.
- Per-client credit budgets, for the agency ceiling.
Questions
What happens to unused credits at the end of the month?
The monthly grant covers its own billing period and does not roll over. Top-up credits are drawn after the monthly grant and stay in the wallet from month to month while your subscription is active. See credits and billing.
Am I charged when a check fails?
No. Credits are reserved before the work starts and released in full if it fails, so a failed check returns your balance to where it was. Refunded work does not appear as spend in the usage report.
Does reading reports or using the MCP read tools cost credits?
No. Reading never costs credits: every report, dashboard, export and read tool on the MCP server is free. Credits are spent only on work that calls an outside provider, such as a check, a crawl, a brief or a draft.
Why was my action refused when the agency wallet has credits?
The client's budget is used up. A per-client budget is a ceiling on the shared agency wallet, so the agency raises that client's allocation to let work continue. See per-client credit budgets.
Last updated