The three rules: credits are spent at the provider's list price with no markup, subscription plans pass through at the provider's own price with no fee added, and our margin is a stated fee rather than a number hidden inside your unit costs. Everything below is the consequence of those three.
We are writing this out because a consolidation service is exactly the kind of product where a margin can be buried invisibly, and the only defence a customer has is arithmetic they can check against a public pricing page.
The numbers
- Free: nothing a month, plus a 15 percent fee taken when you top up credits, with a 50 dollar minimum top-up. Budgets, spend caps and alerts are included; porting in an account you already hold with a provider is not.
- Solo: 19.97 dollars a month. The top-up fee starts at 8 percent and falls as your trailing 30 day API spend grows: 8 percent under 100 dollars, 6.5 percent from 100 to 500, 5 percent from 500 to 2,000, 3.5 percent from 2,000 to 10,000. Port-ins and unlimited projects included.
- Team: per seat, 29.97 dollars for seats 1 to 5, 24.97 dollars for seats 6 to 20, 19.97 dollars from 21 up. The same ladder, one point lower at every step: 7, 5.5, 4 and 3 percent.
- Custom: for accounts past 10,000 dollars a month through the key or 20 seats. 3 percent or less against a commitment, with SSO and SCIM, net terms invoicing and dedicated support.
- Credits: at par. One dollar of credit buys one dollar of provider usage at the provider's list price.
- Provider subscriptions: passed through at the provider's price, on your Squiid invoice, with no fee added.
Why no markup on usage
A markup on usage is the obvious business model for a gateway and we think it is the wrong one, for two reasons.
The first is that it makes the price uncheckable. If we charged 1.12 dollars for a dollar of model usage, you would have no practical way to audit whether the number was 12 percent or 19 percent this month, and you would be pricing your own product on top of a number you cannot verify. Credits at par means you can open the provider's pricing page and check our arithmetic.
The second is that it creates an incentive we do not want. A vendor that earns a percentage of your usage is a vendor that profits when your agent loops. We would rather be paid for the thing we actually do, which is holding the keys, running the gateway, metering the calls and producing one invoice, and that work does not get proportionally harder when your token count doubles.
Where our money comes from
Two places, both visible on your invoice. The monthly platform fee covers the account, the key custody, the gateway, the dashboard, the metering and support. The top-up fee covers payment processing, which is a real cost with a real percentage attached, plus the risk of carrying prepaid balances and the operational cost of provisioning accounts at upstream providers on your behalf.
Taking that fee at top-up rather than per call is deliberate: you pay it once when you add money, rather than on every request, so the cost of a single call is exactly the provider's price and nothing else.
A worked example
A Solo user running a small app with a database on a paid plan and roughly 50 dollars a month of metered usage across hosting, email, storage and a model API:
| Line | Amount | Note |
|---|---|---|
| Squiid Solo | 19.97 | Platform fee |
| Database Pro plan | 25.00 | Passed through at the provider price, no fee |
| Credit top-up | 50.00 | Buys 50.00 of provider usage at list price |
| Top-up fee, 8 percent | 4.00 | Taken once, at purchase, at the starting rate |
| Total | 98.97 | Before tax |
Direct with the providers, the same month is 75 dollars of provider cost spread across several invoices and cards. The difference is 23.97 dollars, and what it buys is one account, one key, one dashboard, one invoice, a balance that pauses at zero instead of overdrawing, and not having to rotate credentials in five consoles when one leaks. Whether that is worth it is your call, and we would rather you could do the subtraction than not.
When Team is worth it, and when it is not
The honest version, because the per seat jump is the question people actually have. Team costs 10 dollars more per person at the first band and lowers the top-up fee by a point at every rung of the ladder, so on fee arithmetic alone it starts paying for itself somewhere around 1,000 dollars of monthly credit spend per seat. Below that, Solo accounts are cheaper than Team seats.
The reason to move earlier is not the percentage, it is ownership. On Solo, the account is a person. The moment a second person needs to deploy, rotate a key or answer a billing question, a personal account becomes a single point of failure, and the handover problem is the one that quietly ends side projects when the person who set everything up stops being available. Team exists so the account belongs to the company rather than to whoever created it.
So: stay on Solo while it is genuinely you, move to Team when someone else needs access or when spend makes the fee difference material. We would rather say that than sell seats to people who do not need them yet.
What we decided not to do
A markup on usage. Covered above. Uncheckable, and the incentives point the wrong way.
A margin on provider subscriptions. Charging 29 dollars for a 25 dollar plan would be easy and nearly invisible. It also means every price on the site becomes a number you have to verify, so we pass them through at cost.
A free tier subsidised by overage. There is a free plan, and it is free of a monthly fee rather than free until you trip a limit. It earns its keep through a higher top-up fee, 15 percent, and a 50 dollar minimum top-up, both visible before you pay. Free tiers that make their money when you exceed them are a trap wearing a friendly face, and they are especially bad for agent workloads that exceed things by accident at 3am.
Annual lock-in for a discount. Maybe later for Custom customers who ask. Not as the default, because most people reading this do not yet know whether the project will exist in six months, and a discount that depends on guessing right about that is not really a discount.
Per service pricing. Charging a little for each service you connect would reward us for the wrong behaviour: you would think twice before adding the ninth service, which is precisely the friction we are supposed to remove. One account covers the whole catalogue.
Charging for the safety features. Spend caps, alerts, pause on zero and the audit log are in every plan. Charging extra for the controls that stop you overspending would be a strange way to earn trust.
The terms, in plain language
Credits expire 12 months after purchase, with warnings before they do. Unused credits are refundable within 24 hours of purchase; the top-up fee is not refundable, because the payment processing cost is already spent. Credits are not transferable between accounts. When the balance hits zero, calls stop being authorised, nothing is deleted, and any subscription that cannot be funded pauses rather than renewing. Auto top-up is optional and has a cap you set. All of that is in the credits and refunds policy in more detail.
If we ever change any of it, the change will be announced rather than discovered, and this page will say what changed and when. That is the whole commitment: you should be able to check our arithmetic against a public pricing page and get the same answer we did.