Nuance
Voice cloning with emotion. 9 engines, zero cloud, all yours. Say it like you mean it.
What is Nuance?
Nuance is a voice cloning and emotion-aware speech synthesis workstation that runs entirely on your machine. Clone a voice from a short sample, then make it say anything — with the emotion you choose, not just the flat robot default.
Nine TTS engines behind one web UI. Each engine has different strengths — some nail the voice similarity, some handle emotion better, some are faster. Nuance lets you switch between them and A/B/C compare with one click.
Emotion-first
Most voice cloning tools give you a voice that sounds like the person but talks like a satnav. Nuance lets you tag emotion clips on the waveform — happy, angry, whispered, sarcastic — and apply them per sentence. The result sounds like someone actually saying those words, not reading them from a teleprompter.
Film mode
Got a screenplay? Nuance reads Fountain format scripts and assigns voices to characters. Each character gets their own cloned voice and emotion profile. Export the whole thing as a multi-character audio file. Table reads just got a lot cheaper.
Why local?
Your voice is biometric data. Nuance never sends it anywhere. All 9 engines run on your GPU (12 GB VRAM minimum, 24 GB recommended). Models download once and run locally forever. No subscriptions, no cloud processing, no “we may use your data to improve our services.”
Status
Alpha. Source-only release. Works but still evolving — expect rough edges, missing UI polish, and the occasional engine that needs a restart. If you’ve got a GPU and some patience, it’s already impressive.
Who it’s for
Content creators who need realistic voiceover without hiring voice actors for every revision. Filmmakers doing pre-viz audio. Podcasters. Game developers. Privacy-conscious users who refuse to upload their voice to someone else’s server. Researchers exploring emotion in synthetic speech.