# Pricing Guide

Real-time usage-based billing (UBB) works differently from traditional SaaS. While conventional billing meters usage and invoices monthly, AI-native products incur costs immediately - with every API call, token generated, or GPU second consumed. This fundamental difference shapes how you should approach pricing.

Credyt supports three pricing models that align with how AI-native products actually deliver and consume value. Each model serves different product strategies and customer expectations.

## Supported Pricing Models[​](#supported-pricing-models "Direct link to Supported Pricing Models")

### 1. Real-time usage-based billing from pre-paid balances[​](#1-real-time-usage-based-billing-from-pre-paid-balances "Direct link to 1. Real-time usage-based billing from pre-paid balances")

Customers fund their wallet upfront with either fiat currency or custom assets (tokens, credits, minutes). Usage draws down from their balance in real time. When balance hits zero, service stops or triggers automatic top-up.

**How it works in Credyt:** Customers top up their wallet through the [Billing Portal](https://docs.credyt.ai/features/billing-portal.md) or via the [Top-up API](https://docs.credyt.ai/features/wallets-and-topups.md). Balances can be held in standard currencies (USD, EUR) or [custom assets](https://docs.credyt.ai/features/assets.md) you define.

As usage occurs, their balance is debited immediately. Customers can enable [auto top-up](https://docs.credyt.ai/features/auto-topup.md) to refill their balances automatically when they fall below a threshold.

**Who uses this model:** OpenAI API, Anthropic Claude API, Replicate, ElevenLabs, and most LLM providers. Common with developer-focused products where customers understand marginal costs.

**When to choose this:**

* You sell tokens, API calls, or compute time
* Infrastructure costs are variable and unpredictable
* Cash flow protection is critical
* You can't afford to front costs for customers
* Your customer base includes developers who expect pay-as-you-go pricing

**Implementation in Credyt:** Define products with `type: usage-based`and`billing_model.type: real_time` in your [product catalog](https://docs.credyt.ai/features/product-catalog.md). Usage events immediately debit the customer's wallet.

For more complex scenarios where pricing varies based on characteristics of the usage (like AI model type, speed, or quality), use [dimensional pricing](https://docs.credyt.ai/features/product-catalog/dimensional-pricing.md) to define different rates for different usage attributes.

See our [Topups Only Model guide](https://docs.credyt.ai/use-cases/top-ups-only-model.md) for further details.

### 2. Recurring fixed fee (subscriptions)[​](#2-recurring-fixed-fee-subscriptions "Direct link to 2. Recurring fixed fee (subscriptions)")

Customers pay a fixed amount on a recurring cycle. This provides predictable baseline revenue while you figure out usage-based pricing.

**How it works in Credyt:** Credyt supports [fixed fee pricing](https://docs.credyt.ai/features/product-catalog/fixed-fee-pricing.md) natively, allowing you to charge recurring subscription fees directly. While Credyt was designed primarily for usage-based billing, recurring fees serve as an ideal starting point when you're still determining how to price usage.

**Why start here:** Beginning with fixed subscriptions lets you generate immediate revenue while building understanding of your cost structure. You can send usage events and associated costs to Credyt without charging for them yet. Using [Profitability analysis](https://docs.credyt.ai/features/profitability.md), you track revenue against actual infrastructure costs per customer, giving you the data needed to confidently transition to usage-based pricing when ready.

**Who uses this model:** Traditional SaaS companies, products prioritizing predictable revenue, or platforms in the early stages of introducing usage-based pricing.

**When to choose this:**

* You need baseline revenue but aren't ready to price usage yet
* You want to understand cost structures before charging for usage
* You're planning to transition to usage-based pricing gradually
* Enterprise customers require predictable billing cycles
* You're still validating product-market fit

**Implementation in Credyt:** Configure products with fixed fee pricing following the [fixed fee pricing guide](https://docs.credyt.ai/features/product-catalog/fixed-fee-pricing.md). Send usage events to track consumption patterns and costs, then use profitability metrics to inform your eventual usage-based pricing model.

### 3. Hybrid billing (subscriptions + entitlements + top-ups)[​](#3-hybrid-billing-subscriptions--entitlements--top-ups "Direct link to 3. Hybrid billing (subscriptions + entitlements + top-ups)")

Base subscription provides predictable revenue and includes usage [entitlements](https://docs.credyt.ai/features/product-catalog/entitlements.md). Additional usage beyond the entitlement can be purchased through top-ups or charged additionally.

**How it works in Credyt:** Combine [fixed fee pricing](https://docs.credyt.ai/features/product-catalog/fixed-fee-pricing.md) with usage-based charges. Subscriptions provide base access and bundled usage allowances through entitlements. When customers exceed their entitlement, they either top up for additional usage or are charged for overages. Both subscription entitlements and top-up credits can coexist in the same wallet.

**Who uses this model:** Clay (seat-based access with credit allocations), Cursor (monthly credit pools per tier), GitHub Copilot (subscriptions with usage caps). Common when you want recurring revenue with usage-based expansion.

**When to choose this:**

* You need baseline recurring revenue plus usage-based growth
* Some features are seat-based, others consumption-based
* Customers want predictability with flexibility for overages
* You're balancing stakeholder preferences for both models
* Different customer segments expect different billing approaches

**Implementation in Credyt:** Configure products with both fixed fees and usage-based pricing. Use [entitlements](https://docs.credyt.ai/features/product-catalog/entitlements.md) to define included usage allocations. Enable top-ups through [Credyt's Billing Portal](https://docs.credyt.ai/features/billing-portal.md) for additional usage.

See our [Hybrid Pricing guide](https://docs.credyt.ai/features/product-catalog/entitlements.md).

## Choosing Your Pricing Model[​](#choosing-your-pricing-model "Direct link to Choosing Your Pricing Model")

Start with three questions:

**1. Can you afford to front infrastructure costs?**

If no: Real-time prepaid credits eliminate cash flow risk. Customers pay before they consume.

If yes: Subscription or hybrid models give customers payment flexibility that enterprise buyers often prefer.

**2. How predictable is customer usage?**

Highly unpredictable: Prepaid credits protect you from runaway costs. Real-time authorization prevents negative balances.

Relatively stable: Subscriptions or hybrid models work well, providing baseline revenue while supporting usage spikes.

**3. What billing model do your customers expect?**

Developers and prosumers: Often prefer pay-as-you-go with real-time visibility into spend.

Enterprise customers: Frequently require invoice cycles and payment terms that align with procurement processes.

## Pricing in Fiat vs Custom Units[​](#pricing-in-fiat-vs-custom-units "Direct link to Pricing in Fiat vs Custom Units")

Credyt supports pricing in both standard currencies (USD, EUR) and [custom assets](https://docs.credyt.ai/features/assets.md) like tokens, minutes, or credits.

**Fiat pricing:** Transparent and enterprise-friendly. Usage deducts directly in currency terms (e.g., $0.006 per API call). Straightforward for reconciliation and familiar to most customers.

**Custom units:** Consumer-friendly for products where abstract units better represent value (e.g., 10 credits per video generation). You define [exchange rates](https://docs.credyt.ai/features/assets.md) between fiat and your custom units. The Billing Portal automatically handles conversions during top-up.

Both approaches can coexist in a single customer wallet. See our [Fiat vs Custom Units guide](https://docs.credyt.ai/use-cases/fiat-vs-custom-units.md) for implementation details.

## Important Note: Invoice-based Usage Billing[​](#important-note-invoice-based-usage-billing "Direct link to Important Note: Invoice-based Usage Billing")

Credyt does not currently support usage-based billing where usage is metered throughout a period and billed in arrears via invoices (similar to AWS Lambda or Twilio). All usage-based pricing in Credyt operates on a prepaid, real-time authorization model.

If your business requires post-facto metering with invoicing after usage occurs, Credyt may not be the right fit for your billing infrastructure.

## Next Steps[​](#next-steps "Direct link to Next Steps")

* Review [Products and Pricing](https://docs.credyt.ai/features/product-catalog.md) to understand how to configure create products and configure pricing
* Read about [Dimensional Pricing](https://docs.credyt.ai/features/product-catalog/dimensional-pricing.md) if your pricing varies by attributes like speed, quality, or model type
* Check our [Quickstart guide](https://docs.credyt.ai/quickstart.md) to start integrating