Set up credit-based billing for an AI platform

This guide walks you through setting up a complete credit-based billing system in Hyperline, where customers purchase credits and consume them at different rates depending on which AI model they use.

Context

You run an AI platform that offers access to multiple LLM models with different capabilities and costs. Rather than billing per token for each model separately, you want to sell credit packs that customers can spend across all models — with more powerful models consuming more credits per usage. Your platform offers three model tiers:

Model	Use case	Credits per 1,000 tokens
`nova-fast`	Quick tasks, high volume	1
`nova-pro`	Balanced performance	4
`nova-ultra`	Complex reasoning	15

A customer with 10,000 credits could use them flexibly: heavy usage of the budget model, occasional use of the premium model, or any combination.

Prerequisites

Before you begin, make sure you have:

Access to your Hyperline account (we recommend using test mode for this setup)
An API key to send usage events
A clear understanding of the models you want to bill and their relative cost ratios

When signed up and landed on Hyperline, you can use the test mode to experiment before going live.

Steps

1. Send usage events via the API

Every time a customer uses one of your models, your backend should send a billable event to Hyperline with the model name and the number of tokens consumed.

curl -X POST https://ingest.hyperline.co/v1/events \
  -H "Authorization: Bearer <API key>" \
  -H "Content-Type: application/json" \
  -d '{
    "events": [
      {
        "customer_id": "cus_xxxxxxxxxxxxx",
        "event_type": "llm_token_usage",
        "timestamp": "2026-03-25T14:30:00Z",
        "properties": {
          "model": "nova-pro",
          "tokens": 1500
        }
      }
    ]
  }'

Each event should include:

customer_id: the Hyperline customer ID (or external_id if you’ve mapped your internal IDs)
event_type: a consistent name for your token usage events (e.g., llm_token_usage)
properties.model: the model identifier used for the request
properties.tokens: the total number of tokens consumed (input + output)

For testing, you can skip this step and use simulated events directly in test mode to generate fake usage data.

Learn more about event ingestion in our send usage data guide.

2. Create one aggregator per model

Each model tier needs its own aggregator so that Hyperline can track consumption separately and apply a different credit weight to each. Go to Usage > Aggregators and create three aggregators:

Aggregator name	Event type	Operation	Property	Filter
`nova-fast tokens`	`llm_token_usage`	Sum	`tokens`	`model` = `nova-fast`
`nova-pro tokens`	`llm_token_usage`	Sum	`tokens`	`model` = `nova-pro`
`nova-ultra tokens`	`llm_token_usage`	Sum	`tokens`	`model` = `nova-ultra`

All three aggregators listen to the same llm_token_usage event type but use a filter on the model property to isolate each model’s consumption. The Sum operation on tokens gives you the total number of tokens consumed per model.

Example 💡

When a customer sends 3 requests to nova-pro consuming 1,200, 800, and 2,000 tokens respectively, the nova-pro tokens aggregator reports a total of 4,000 tokens for that period.

3. Create a credit product with weighted aggregators

Now create a credit product that ties all three aggregators together with their respective weights.

Create a new credit product

Go to Products, click New product, and select Credit.

Configure the aggregators and weights

In the Aggregators section, add the three aggregators you just created and set the weight for each:

Aggregator	Weight
`nova-fast tokens`	1
`nova-pro tokens`	4
`nova-ultra tokens`	15

The weight determines how many credits are consumed per unit of usage. With these weights, 1,000 tokens on nova-ultra consumes 15 credits, while the same 1,000 tokens on nova-fast consumes only 1 credit.

Set up bundle pricing

In the Prices section, configure your credit packs. For example:

Bundle size	Price
10,000 credits	$100
50,000 credits	$400
200,000 credits	$1,200

These bundles will be available for purchase via the customer portal and auto-topup.

Save the product

Click Save to add the credit product to your catalog.

For more details on multi-aggregator credit products, see the credits overview.

4. Set up a credit balance on a customer

With your credit product ready, enable a credit balance on a customer.

Navigate to the customer's Credits tab

Go to Customers, select a customer, and click the Credits tab.

Add a credit balance

Click Add credit balance and select the credit product you created.Configure the initial settings:

Starting balance: the number of credits to grant immediately (e.g., 10,000)
Warning limit: the threshold at which a low-balance warning appears on the customer portal (e.g., 1,000 credits)

5. Add the credit product to a subscription

To automate credit replenishment and invoicing, add the credit product to a subscription.

Create or update a subscription

From the customer’s page, click Actions > Assign new subscription, or open an existing subscription.

Add the credit product

Click Add product, select your credit product, and choose a bundle pricing tier.

Configure auto-topup (optional)

To ensure your customer never runs out of credits, configure auto-topup on their credit balance:

Threshold: when the balance drops below this number, auto-topup triggers (e.g., 500 credits)
Amount: how many credits to add (e.g., 10,000)
Pricing: select a bundle price or set a custom amount

Auto-topup automatically creates an invoice, charges the customer’s payment method, and adds credits to their balance.

Auto-topup requires the customer to have a valid payment method on file. If the payment fails, the topup will not be processed.

6. Monitor consumption

Once the customer starts using your models, you can track credit consumption from the customer’s Credits tab. The credits chart shows balance evolution over time. Use the date range and granularity controls (day, week, month) to analyze consumption patterns.

Click All transactions to view the full transactions table, which lists every credit movement — top-ups, usage deductions, and expirations. For multi-aggregator products, an Aggregator column shows which model triggered each usage transaction. Click Explore on any usage transaction to inspect and export the underlying consumption events.

Example 💡

A customer starts the month with 10,000 credits. They make:

2,000 nova-fast calls averaging 1,000 tokens each → 2,000 x 1 = 2,000 credits
300 nova-pro calls averaging 1,500 tokens each → 450 x 4 = 1,800 credits
50 nova-ultra calls averaging 2,000 tokens each → 100 x 15 = 1,500 credits

Total consumption: 5,300 credits. Remaining balance: 4,700 credits.The transactions table shows each deduction tagged with the corresponding aggregator, so you can see exactly which model is driving consumption.

Next steps

Credits overview

Learn more about credit products and configuration options

Credit expiration

Configure credits to expire after a set period or at the end of each billing cycle

Customer portal

Let customers view their balance and purchase credits from the portal

Credits API

Manage credits programmatically via the API

Is something still unclear? Don’t hesitate to reach out to our team via the in-app chat if you need additional support.

Implement Hyperline

CPQ

Set up credit-based billing for an AI platform

Context

Prerequisites