Providers

StepFun

STT TTS S2S 7 models 7 voices based in China self-serve API data from documentation synced 2026-11-16

StepFun (阶跃星辰) is a Shanghai AI lab founded in 2023. Its StepAudio models cover speech recognition, text-to-speech and realtime speech-to-speech, sold pay-as-you-go through the international StepFun platform.

Speech-to-text 3

CompareModelLangsFeaturesPriceStatus
stepaudio-3-asr-maxStepAudio 3 ASR Max 6 0/11 $0.60 per hour Default
stepaudio-2.5-asrStepAudio 2.5 ASR 6 0/11 $0.05 per hour
stepaudio-2.5-asr-streamStepAudio 2.5 ASR Stream 6 1/11 $0.40 per hour

Text-to-speech 2

CompareModelLangsFeaturesPriceStatus
stepaudio-3-ttsStepAudio 3 TTS 6 4/8 $14.00 per 1M chars Default
stepaudio-2.5-ttsStepAudio 2.5 TTS 6 3/8 $85.00 per 1M chars

Speech-to-speech 2

CompareModelLangsFeaturesPriceStatus
stepaudio-2.5-realtimeStepAudio 2.5 Realtime 1 0/19 $1.50 per 1M tokens Default
stepaudio-3-realtime-previewStepAudio 3 Realtime (Preview)Beta 2 0/19 – Beta

Voices 7

Browse and filter all of them.

Esc

↑↓ move↵ open