Credits were previously referred to as "characters." We updated this terminology to streamline pricing and make it easier to understand, though the value remains unchanged.
Previously, this terminology was confusing, as “characters” were also used for tools that had non-text-related inputs like speech-to-speech and dubbing. We think "credits" is a more appropriate name as they can be used across all of the tools in our product.
Regarding text-to-speech, you will still see the terminology "characters" as this refers to written text characters. For our English V1, Multilingual V1, and Multilingual V2 models, each text "character" you generate into speech will consume 1 credit. For our Turbo v2 and v2.5 models, each text "character" will consume 0.5 credits, unless you have a custom enterprise plan. Note that some shared voices in our Voice Library have a custom rate that applies a multiplier to your credit consumption.