Description
Dubbing a novel, audiobook or video, the giveaway is always the machine: flat tone, odd phrasing, no breath, none of the ums and pauses real speakers leave in. MiniMax Speech goes after exactly that: strong Chinese performance, voice cloning and long-form text to speech, and version 2.8 adds native sound tags that model breaths, laughs and spoken fillers.
There is a web app you can try right away: type text, pick a voice, generate. It also has a voice library, voice design, voice cloning and a voice isolator, plus an API for building it into your own product.
Long-form narration: Built for novels, audiobooks and video voice-over
Sound tags: Version 2.8 expresses breaths, laughs and fillers with tags, keeping natural rhythm and pauses
10-second cloning: A sample of about ten seconds captures a speaker's timbre and breathiness
Voice library and design: Browse and audition ready voices, or design a new one from a description
Voice isolator: Separates speech from background noise to cut noise and digital artifacts
API: An API for building speech synthesis into your own apps
There is a web app you can try right away: type text, pick a voice, generate. It also has a voice library, voice design, voice cloning and a voice isolator, plus an API for building it into your own product.
Features
Long-form narration: Built for novels, audiobooks and video voice-over
Sound tags: Version 2.8 expresses breaths, laughs and fillers with tags, keeping natural rhythm and pauses
10-second cloning: A sample of about ten seconds captures a speaker's timbre and breathiness
Voice library and design: Browse and audition ready voices, or design a new one from a description
Voice isolator: Separates speech from background noise to cut noise and digital artifacts
API: An API for building speech synthesis into your own apps
