Imagine a world where creators have more voices for their projects, and voice actors get paid a fair share.
(And while we're pipe-dreaming, let's make it a world where TTS is an automated accessibility tool, not an audio workflow for creative projects...)
While the AI world focuses on text-to-speech - the alchemy project that aims to take flat text and turn it into anything other than a flat voice - we're spotlighting speech-to-speech. TTS may be a quick solution for some small-scale projects, but when love goes into developing characters and worlds, love needs to come out of the voice that brings them into reality.
Starts from flat text and tries to conjure a performance that was never given. The emotion is guessed, or painstakingly added using emotional tags, not acted.
Demands the qualities of a voice actor - emotion, pacing, volume, exaggeration, accent, everything that gives characters life. In return it offers a richer experience for the listener: one that suspends disbelief while still being brought from one human to another.
Compare a real human, STS generated from their performances, and the same line generated by the same voice model in TTS. Drag the slider to morph between the voices, and hear the same pacing, same emphasis, same breaths, and even the same mistakes in a different voice.
We believe that by building a community library of actors and their voices, this brilliant creative technology can be utilised and improved by the people who care about it most.




* Every original take above was performed by a software developer, not a voice actor, so we might only imagine what STS could do with actual talent. Also, wouldn't it be wonderful if all the authors of these voice models had been properly credited...?
We aim to build a community of voice actors who are credited and paid for the use of their voice, offering real opportunities for earning a passive income from a wider range of creators who love their voices, but couldn't ordinarily afford to hire a wide variety of voice actors for smaller projects.
By offering a performative, human-centric end product that makes even the most advanced text to speech sound like Microsoft Sam, we're hoping to kick copy-paste-gen mentality out of the modern creative pipeline for good.
Every voice carries its actor's name, and every use rewards them. No uncredited, underpaid voices going stale in a library while offering no exposure to the actors that put hours into perfecting the training models.
Get your voice out there, then keep earning as creators make your tone, timbre and tenor a part of their character. Be ready for specialist projects as you become a credited voice that can be contacted for studio work where STS doesn't cut it.
Use your acting talent free of the restriction of your vocal pitch or style - act out more than a single character off the back of your spoken personality, and make your talent more desirable in the modern market.
ElevenLabs' APIs are currently included in Narrative Studio as a proof of concept for speech-to-speech. While it shows what STS can do, voices remain uncredited and the contributing actors are not meaningfully rewarded. Let's build a better solution.
Our final goal is to move to a community-driven STS that offers the best of all worlds to all parties: creators get the voices they need, and the actors behind them are named and paid.
Optional third-party STS, wired in to prove the workflow.
A voice library built with real actors to make their talent known to a broad audience that might otherwise never hear of them.
Whether you are a voice actor or someone who needs them, or someone who is stuck with a TTS workflow, we'd love to build this with you. Try STS for yourself, and see if you agree that it's the way to a better creative future for us all.