Credit-Based Pricing

Credit-based pricing charges customers in a virtual currency - credits - that they buy in advance and spend as they use the product, with different actions consuming different amounts. It's become the default model for AI tools, where one image generation, video render, or model call costs a set number of credits. The appeal is flexibility: a single credit pool can span wildly different features whose underlying costs differ, letting you price by value and absorb variable model costs without exposing raw per-token billing to customers. Credits also create useful psychology - prepaid balances feel spent already, encouraging usage, and running low nudges a top-up. The risks are real: opaque credit-to-value mapping frustrates buyers who can't predict what a task will cost, and credits that expire or don't roll over can feel punitive. If you use this model, publish a clear credit calculator, align credit costs to your actual COGS so heavy usage stays profitable, and be transparent about expiry and overage so trust survives the first surprise invoice.