This post records the Vois 1.x voice library. It is kept so older projects and references remain understandable, not as a catalog for a new production.
If you are choosing a voice today, start with the current Vois voice library. Vois now has 100+ voices across 21 categories, with a browser for testing your own material before you commit. The legacy labels below may not match the current library, so use them as historical context rather than production instructions.
What the original library was built to solve
The old library grouped a smaller set of voices by language and broad role. The practical problem was the same as it is now: a narrator that sounds good for thirty seconds can be tiring across an audiobook chapter, while a crisp tutorial voice may be wrong for a warm story.
Today's Vois workflow makes that decision easier. Paste a representative passage, preview a few candidates, and listen for clarity, pace, and emotional fit. Treat the script and voice as a pair, because changing either one can change how a line lands.
American English in the original collection
American English was the largest family in Vois 1.x. Historical examples included af_heart, a warm storytelling option, and af_nova, a brighter conversational voice. am_adam was often chosen for measured tutorials, while af_alloy and am_echo were used for more formal or educational reads.
The lesson still holds: choose for the job, not for a name. For a tutorial, test terminology and transitions. For a podcast, test a full conversational segment. For a story, test an emotional passage and a quiet one. The current browser lets you make that call with the actual copy rather than a description alone.
British English in the original collection
The historical British options included bf_emma, bf_alice, bf_lily, bm_daniel, bm_george, and bm_lewis. They offered a range from warm storytelling to documentary-style delivery.
Accent is only one part of casting. A documentary may need restraint, while a character-led series may need more color. Preview a longer extract at the listening pace you expect, then choose the voice that stays comfortable and clear.
What changed for multilingual production
Vois 1.x listed dedicated options for Japanese, Mandarin, Spanish, French, Hindi, Italian, and Brazilian Portuguese. That mattered because translated copy read in an unsuitable voice can sound disconnected from its intended listeners.
The current Vois workflow is broader. Pro users can use multilingual voices across 600+ languages, but translation and pronunciation still need human review. A familiar voice identity is useful only when the words, cultural context, and pacing are right for the people hearing them.
How to choose a current Vois voice
Start with the listening context:
- Podcasts: audition a full exchange or a few minutes of exposition, not an isolated tagline.
- Audiobooks: test both an emotional scene and explanatory prose. The narrator needs to stay listenable for hours.
- Tutorials: prioritize intelligibility and measured pacing over personality.
- Localization: use reviewed translations and audition language-specific passages before publishing.
Save the short list in the project, use the same test passage for each candidate, and get feedback from people who match the intended listener. That creates a repeatable casting decision instead of a guess.
Voice cloning when the library is not enough
If the project needs a voice you own or have permission to use, Vois supports voice cloning from a clear short sample. Keep consent explicit, use a sample that represents the delivery you need, and audition the clone before it reaches a final script.
Cloning is most useful when consistency is part of the project, such as a recurring host, approved brand speaker, or character. It is not a reason to skip permission or casting review.
Why the studio workflow matters
Current Vois voices can be previewed with your text, adjusted from 0.5x to 2.0x, and used in a local script-to-export workflow. That gives you room to review one sentence, regenerate only the section that needs work, and keep a series consistent without treating a voice selection as permanent.
This legacy guide explains where the library started. For a new project, explore the current voices, then Get started with a real passage from your script.
The Vois Team