Selling tokens? Stop each request before the balance runs out.
PortaAIM checks prepaid balances in the request path and rates every model you serve. It runs in your country, and comes with the source code.
A 30-minute technical call. We reply within two business days.
Currently in pilots with early sovereign AI and GPU cloud operators · 25 years of real-time billing · 500+ operators in nearly 100 countries
Where does token billing break?
118 of the 156 AI platforms we reviewed sell mainly on prepaid credits or a wallet.
- Without PortaAIM
Usage gets counted after the response
Many usage billing tools record the request first and subtract the cost after. That works for invoices, not for prepaid balances.
- A request costing $1 can still go through with a balance of just 5 cents
- Every new model or unit means a billing code change
- Bundling tokens with GPU rental needs custom code
- Peak rates, quotas, or regional prices need custom billing code
- Resellers in other markets get settled in spreadsheets
- With PortaAIM
PortaAIM checks the balance before the request runs
PortaAIM sits in the request path, checking the balance, applying rates for every model and unit, and billing customers and resellers from one engine.
- Check the balance and reserve the cost before the request runs
- Rate tokens, tool calls, characters, and GPU hours without schema changes
- Bundle tokens, credits, and GPU rental in one plan
- Peak and off-peak rates, quotas and regional pricing, without custom code
- Settle resellers in the same system as your customers
Built for token sellers, with or without your own GPUs
Inference providers and GPU clouds expanding into inference choose PortaAIM for different reasons.
Inference providers and sovereign AI clouds
- Runs in your country or your own environment
- Source code is provided as standard
- Manage prepaid balances in one central ledger instead of separate balances per site
GPU clouds selling inference
- Add a per-token catalogue next to GPU hours, without a second billing system
- Manage metered usage and flat GPU rental in one system and one portal
- Offer committed-use and minimum-commitment pricing next to usage
- Connect your console and systems through full API and MCP access
- Want to see your own pricing running on PortaAIM?
How PortaAIM charges a request
PortaAIM sits in the request path, between your gateway and the model.
A request arrives at your AI gateway
PortaAIM checks the balance and reserves the cost*
The request runs if there are sufficient funds, otherwise it's blocked
Actual usage is calculated, and the remaining amount is released
Charges show on the invoice, and resellers are settled
* Choose how strict the check is, separately for each customer:
- Full check: reserve the worst-case cost before the request runs
- Light check: reserve less, then add more as the response runs
- No check: charge what was used, for postpaid or trusted customers
From first call to a demo on your pricing
One call, a demo on your own pricing, then you decide.
Technical call, 30 minutes
A PortaOne engineer walks through what you sell, how you charge, and where your billing needs to run.
Demo on your own pricing
We set up a demo using your data within a week.
Your decision
You decide whether to take it further. No commitment before that.
Is PortaAIM a fit for you?
You sell tokens, GPU hours, or credit allowances to many customers, and:
You sell several products and want them billed from one system
Your prices differ by region and by customer
You sell usage, subscriptions, and allowances at the same time
You need the balance checked at the request, not at the end of the month
You add new billable units faster than you can ship releases
Talk to an engineer about your billing
Tell us what you sell and where it must run, and one of our engineers will reply within two business days.
A 30-minute technical call. No commitment.
FAQ
Does PortaAIM stop a request when the prepaid balance is too low?
Yes, if you choose a full check. PortaAIM checks the balance before the request runs, reserves the worst-case cost, and blocks the request if the balance can’t cover it. For postpaid or trusted customers, you can skip the check and charge what was used.
We already use a usage billing tool. What changes?
PortaAIM checks the balance in the request path before the model runs, instead of subtracting usage after the response.
Do we have to replace our AI gateway?
No. Your gateway calls the PortaAIM API to check the balance before each request and report usage after it. Your engineers add that call.
Can PortaAIM run entirely on our own infrastructure?
Yes. PortaAIM runs on-premise or in-country, with source code supplied as standard. Usage is tracked and billed locally, so nothing leaves your environment.
Can we settle resellers in other markets?
Yes. PortaAIM settles resellers in the same system, whether the capacity is yours or bought from another provider.
What support is included with PortaAIM, given that we'd be relying on it for our revenue?
PortaOne billing engineers support PortaAIM 24/7, with response commitments in your contract.
About PortaOne
Since 2001, PortaOne has been developing billing, business support systems, and cloud telephony software, helping telcos connect people.
Our journey has seen the rise of over 500 forward-thinking telcos in more than 100 countries, aiding them in becoming market leaders while keeping the total cost of ownership for their business support systems under control.
Our customers speak to our commitment, citing our support as the best in the industry. We offer a lifetime license without limits and an open-architecture platform, enabling swift in-house development of new services. Access to source code is granted, ensuring transparency and security validation.
