Step-Audio-EditX
StepFun's LLM-grade audio editing — emotion, tone, style — "Photoshop for audio"
StepFun's LLM-grade audio editing model with editable emotion, tone, style, and paralinguistic features — "Photoshop for audio".
Voice AI: lifelike TTS, accurate ASR, and real-time voice agents.
Key features
- 3B-parameter RL audio model for expressive iterative audio editing
- Edits emotion and speaking style with dozens of options
- Paralinguistic editing with 10 tags like breathing and laughter
- Zero-shot TTS for Mandarin, English, Sichuanese, Cantonese, Japanese, Korean
Good for
- Audio editors iteratively changing emotion or style of speech
- Developers adding polyphone pronunciation control for Mandarin
Stack & tags
vLLMaudio-editingttsemotion-editingparalinguisticzero-shotmultilingual