Talent
Your voice library. Every voice you've built or borrowed.
Train a custom AI voice from your own recording, or import an existing RVC model. Every voice gets a catalog number, a consent record, and a resume that fills in automatically as it's cast across the suite.
Not yet available.
Cast
Talent is where a voice is born — trained from a real recording or brought in from elsewhere.
A Recording, or a Model
10–30 minutes of source audio to train from scratch, or an existing .pth model to import directly.
One Shared Library
Every voice you build or import is instantly available to TableRead and Studio — cast once, use anywhere in the suite.
Two ways to get a voice into your library.
Train one from scratch, or bring one you already have. Either way, it lands in the same library, ready to cast.
- Train from a Recording — record or import roughly 10–30 minutes of source audio, name it, and train
- Import an Existing Model — point at an existing RVC .pth (and optional .index) file and register it directly, no training required
- Real Preview — hear an actual converted sample of a freshly trained voice, not a placeholder
- Shared Everywhere — every voice is instantly available in TableRead and Studio, no export/import step between apps
Consent and quality control, built into the model.
Consent Tracking
A consent attestation is recorded alongside every trained voice — because training a voice model from someone's recording is a real responsibility, not just a technical step.
Catalog Numbers
Every voice gets a permanent, human-friendly catalog number — easy to reference, easy to keep straight across a growing library.
Testing vs. Ready for Casting
Flag a voice as still-testing while you evaluate it, then mark it Ready for Casting once you trust it for real production work.
On-Device, Always
Training and conversion run entirely on your Mac. No recording, no voice model, is ever uploaded anywhere.
A resume that fills itself in.
Every voice keeps a running record of which projects and characters it's performed — credited automatically the moment TableRead generates a read-through or Studio exports a finished take. No spreadsheet, no manual bookkeeping.
Simple, honest pricing.
Free Trial
- Full training & import — no limits
- Consent tracking & catalog numbers
- Testing / Ready for Casting status
- Voice resume & credits
- Shared library across the suite
Talent
- Everything in the trial
- Buy once · yours forever
- Free updates
- macOS 14 Sonoma or later
Talent is the deepest tech in the suite. Get it and 4 more apps in the PolySpeak bundle for $129 — save $61 versus buying separately.
Frequently asked questions
How much recording do I need to train a voice?
Roughly 10 to 30 minutes of clean source audio. Record it directly or import an existing recording — either way, Talent walks you through naming the voice and setting how many training epochs to run.
Can I use a voice model I already trained elsewhere?
Yes. Talent can import an existing RVC .pth model file (with an optional .index file) directly, no retraining required — it's registered into your library the same way a freshly trained voice would be.
Do I need consent to train a voice from someone's recording?
Yes, and it's your responsibility as the user to have it, or use only your own voice. Talent records a consent attestation alongside every trained voice, but DataIsland cannot verify real-world consent — see our Terms of Use for the full policy.
Does any of my recorded audio leave my Mac?
No. Voice training and conversion run entirely on-device through a local pipeline. Nothing is uploaded to DataIsland or any third party at any point.
Do I need TableRead or Studio to use Talent?
No — Talent is a complete, standalone voice library on its own. TableRead and Studio become more useful once you have voices to cast, but training and managing a library works fully without either app installed.
Is Talent included in the PolySpeak bundle?
Yes. The PolySpeak bundle includes Talent, PolyScript, TableRead, Studio, and PolyProject for $129 one-time — a savings of $61 versus the $189.95 combined individual price. See the full breakdown on the PolySpeak bundle page.