The most accurate transcription model ever - supporting 90+ languages with exceptional accuracy on long, complex recordings across diverse speakers.
Supports 90+ languages. Set the primary language of the uploaded file, or By default, it’s detected automatically.
Cinematic audio from a single prompt. Ads, stories, songs, and music scores. Voice, ambience, and effects, all generated together.
Reading file…
Fourteen starting points across ads, dialogue, songs, scores, and more. Pick one and start generating.
From repeatable production runs to super creative experiments, choose the model that fits your workflow.
The best of both worlds. Pulse-level consistency with Mercury-level creativity.x
Built for stories and audio ads. The same prompt delivers the same audio, every time.
Built for songs, singing, and music creation. Fast, creative scores and immersive results every run.
Everything you need to know about creating with SuperTTS.
SuperTTS turns your written prompts into natural, scene-aware audio using advanced AI voice models. Write a direction, pick a model, and generate narration, ads, and voice content without recording.
SuperSeed supports languages optimized for dialogue-heavy production. Pulse and Mercury support a broader language set. All models are best suited for short-form audio under 3 minutes.
You can export generated audio in MP3 and other common formats. Choose your sample rate in Audio Settings before you generate.
Yes. Each SuperTTS prompt supports up to 3,000 characters per generation. Use duration and variation controls to fine-tune the result.
Yes. Use the sidebar audio controls to adjust speed, volume, and pitch so the output matches your scene or brand style.
Use Pulse 1.0 for stories and audio ads when you need repeatable, production-ready consistency with custom voice, SFX, and BGM placement. Choose SuperSeed 1.0 for music — scores, soundscapes, and creative mixed audio with expressive variation. Choose Mercury 1.0 for songs, singing, and music BGM — fast, immersive results with a fresh take every run.
SuperSeed and Mercury typically generate quickly. Pulse prioritizes consistency and may take longer on complex prompts.
Pulse 1.0 works with all explicit prompts. On SuperSeed 1.0 and Mercury 1.0, some prompts may not work. Subject to our terms of service.
Generated files are saved to your account history so you can replay, download, or reuse them. You can delete items anytime from your SuperTTS library.
Yes. Audio created with your subscription can be used commercially according to your plan terms, including ads, podcasts, products, and client work.
Ready when you are
~75ms
ProVoices
Discover powerful features that give you total creative control from cloning your voice to fine-tuning every detail.
Helping teams and businesses across industries bring content to life.
Billions of words. Millions of voices. Thousands of businesses.
Growing every minute.
2B+
Audio files generated to date
200M+
Credits converted to voice every single day
5M+
Registered users across 120+ countries
20k+
Businesses power their workflows with our voice API
