How to stop repeat credit abuse in your SaaS
Repeat credit abuse never spikes, so nobody looks. How refreshes and signup credits leak money, which basic fixes break, and what to ask at every grant.
It's 7:41 a.m. on the 27th, and a customer on your $29 plan has sent you a message: "Every generation has failed since 6. We've got a client demo at 11. What's going on?"
You open your model provider's dashboard. The $3,500 monthly limit you set was reached at 6:12 a.m., so for the last 89 minutes your product has been returning errors to everyone, paying customers included. You raise the limit from your phone, apologize and refund that customer's month. Then you go looking for whatever ate the budget, and you can't find it. There's no spike. This month looks like last month, which looked like the month before.
Stay with this scenario, because the arithmetic is where it hurts. Your free plan refreshes 1,000 credits a month, and a fully used account costs you $1.50 a month. One person running 40 free accounts takes $60 a month (40 × $1.50). That's more than twice what the customer with the client demo pays you. Now say 25 people are doing the same thing:
- 25 people × 40 accounts = 1,000 accounts
- 1,000 accounts × $1.50 = $1,500 a month
- $1,500 × 12 months = $18,000 a year
Those 1,000 accounts don't stand out. Every one has a verified email, every one is under its limit, and the credits come back on their own each month. That's why how to stop repeat credit abuse is a different question from how to survive an overnight attack. An attack wakes someone up. This just becomes part of what your business costs.
Without those accounts, your spend on the morning of the 27th would have been $2,000 ($3,500 minus $1,500), and that demo wouldn't have been at risk. $1,500 a month is also about what 52 customers on your $29 plan pay you (52 × $29 = $1,508), before you've paid to serve any of them. If your company spends $15,000 a month, $18,000 a year is 1.2 months of runway handed to 25 people who were never going to buy. On the 1st of next month, it starts again.
So you reach for the fix you can ship today, and you cut the free plan from 1,000 credits to 200. The bill comes down. Then the replies come in. Someone who's been building a side project on your free plan for months asks what happened. Someone who was close to upgrading stops logging in and never says why. And the 25 people have an easy way back: 1,000 ÷ 200 = 5, so the same credits now take 200 accounts instead of 40. Your honest users got a fifth of the plan they signed up for because of what 25 people did, and those 25 are the only ones who can get the old amount back.
Why repeat credit abuse stays invisible
Credits get handed out per account, so they get abused per account. A signup bonus pays out once for each new account. A monthly refresh pays out 12 times a year, and a daily allowance pays out 365 times. Multiply by 40 accounts and one person collects 480 monthly grants a year, or 14,600 daily ones.
No single grant is wrong, which is why nothing flags it. Your limits count usage inside an account, and this spreads usage across accounts. The cost lands in the same line as every real user's, your budget grows around it, and your growth chart counts those 1,000 accounts as users. In an AI app this is the pattern behind credit farming, and it's worth working out what a free user really costs you before you pick a fix. If your costs jumped overnight instead, you want the first-hour runbook.
The obvious fix is a rule: one grant per account per period. Build it, because it's worth having. Just know how far it reaches.
Start with a grant ledger
A ledger records every grant before any credits move, so the database is what refuses a second one. This version is for PostgreSQL:
create table credit_grants (
account_id uuid not null,
grant_period text not null,
credits integer not null,
granted_at timestamptz not null default now(),
primary key (account_id, grant_period)
);
insert into credit_grants (account_id, grant_period, credits)
values (:account_id, :grant_period, :credits)
on conflict (account_id, grant_period) do nothing
returning account_id;grant_period is a value you define for each allowance: signup for a one-time bonus, 2026-11 for a monthly refresh, 2026-11-08 for a daily one. Add credits to the balance only when the insert returns a row, and do both in one transaction. A retried job, a scheduler that fires twice or two requests landing together now produce one grant.
What the ledger can't do
It stops double grants within one account. It does nothing about a new account. A new account_id has no rows, so the insert succeeds and the full allowance goes out.
The basic checks, and how each one breaks
Most teams try these next, and each one does catch something. Start with email, where the limits are easiest to see.
A unique-email rule compares strings, and one mailbox can have many strings. Gmail ignores dots, so jane.doe@gmail.com and janedoe@gmail.com reach the same inbox. Plus addressing is similar: mail to jane+anything@gmail.com lands with Jane too. That's simply how email works.
| Basic fix | What it catches | What it misses | Who it wrongly blocks | What it costs to keep working |
|---|---|---|---|---|
| Email verification | Typos and dead addresses | Any real second inbox | People whose code lands in spam | Sending and deliverability work |
| Unique emails | The same address twice | Every different address | Almost nobody | Very little |
| Disposable-domain blocks | Known throwaway services | New domains and ordinary mailboxes | People using privacy forwarding | A list someone updates forever |
| Gmail dots and plus addressing | One Gmail inbox written several ways | A second real mailbox | People who sort their mail with plus addresses | Upkeep as mail services change |
| CAPTCHA | Simple scripts | People signing up by hand | People using screen readers or privacy tools | Fees and lost signups |
| Phone requirement | Casual second accounts | Anyone with another number | People without a reliable mobile number | A charge for every text |
| Card requirement | Casual second accounts | Anyone with another card | Students and people who want to try first | Fewer signups |
| One per IP address | Repeat signups on one connection | Any other network | Offices, campuses and households | Exceptions and tickets |
| VPN blocks | Known VPN ranges | Everything off your list | Privacy-minded people, staff on company VPNs | Another list to maintain |
| Rate limits | Bursts from one account | 40 accounts under the limit | Real users in a busy hour | Tuning per endpoint |
| Usage caps | One account overrunning | Each new account's fresh cap | Anyone at the cap | Deciding when your product stops |
| Smaller allowances or shorter trials | Some value per account | More accounts make up the difference | Every honest user | Fewer people reach paying |
| Manual review | What a person can spot | Anything after hours or spread out slowly | Whoever looks odd to a tired reviewer | Hours every month |
Say you add a disposable-domain block this week. A few days later the new accounts use a domain that isn't on your list, and a paying customer's colleague can't sign up because their forwarding service is. The 1,000 accounts you already have keep refreshing anyway, because a signup check only runs at signup.
Stack them all and your form asks for a verified email, a puzzle, a phone number and a card. Manual review on top, at one minute per account, is 1,000 minutes, nearly 17 hours, and it's due again next month. Your honest users feel every one of those steps, and CAPTCHA has limits of its own. The cost of blocking real people is the part that never shows up on an invoice.
Ask one question at every grant
Every fix above asks about the signup. Is this email real? Is this a human? What you need to know comes later, when credits are about to move: has this person already had this benefit?
List every grant. Signup credits, refreshes, daily allowances and top-ups are each their own benefit.
Write down each limit. Signup credits might be once per person. A refresh might be once per person per month.
Send every grant through one server function that writes the ledger. Grant a refresh when someone comes back, so dormant accounts cost nothing.
Give free traffic its own budget and a hard cap, so a bad month on the free plan can't lock out paying customers.
Decide what a no looks like. Offer a paid plan and a way to reach you, because some of the people you turn away will be real.
Your records can answer that question for an account in one query. They can't answer it for a person. A new account arrives with nothing on paper tying it to the last one, while a real second person on the same office connection can look just as suspicious. Getting that right on every grant, and keeping it right, is a job of its own.
Evaluating the claim without building it yourself
You don't have to build your own abuse prevention system. AbuseZero is built to evaluate whether each claim looks like a repeat, and it starts where every row in that table ran out. You get multi-layer AI risk analysis that examines 140,000+ signals, detects linked claims and flags abuse that single checks miss. Signals are read together, never one at a time. AbuseZero uses a proprietary algorithm, with AI that checks the claim again. AbuseZero recognizes repeat devices across new accounts, even when email addresses keep changing, and detects repeated payment fingerprints across users, even when new accounts and emails hide the connection. It checks mailbox aliases and disposable domains too. Creating a fresh account isn't enough to erase previous activity.
You make one API call before each grant and get a clear allow or deny, under rules you control. Eligibility can be once per user, or once per calendar day, week or month, or once per rolling number of days, so your signup credits and your monthly refresh can each have the right limit. Observe mode shows you what would be denied before anyone is blocked, and the quickstart walks you through your first benefit.
Frequently asked questions
What is repeat credit abuse?
It's one person collecting the same credits more times than you meant to give them, usually by running several accounts.
Should I cut my free plan to stop repeat credit abuse?
Keep that as a last resort. A smaller allowance lowers your cost per account, but every honest user loses credits and someone running many accounts can open more. See how to stop free plan abuse in a freemium SaaS. AbuseZero evaluates each claim at the moment of the grant, so the size of your plan can stay a product decision.
Is every second account abuse?
No. People have work and personal accounts, and families share a connection. Write down what you mean to give each person.
Keep your free plan for the people it's for
Start with Builder
Pay as you go. Protect your first benefit today.
See pricing
Builder, Pro, Scale and Enterprise.
Talk to us about Enterprise
High-volume credits and custom terms.