A voice can sound right in a demo and still leave you with a lot of work to finish a project. That's the gap I care about: the distance between hearing something promising and feeling confident enough to send the finished audio.
The next Vois release is being built around that concern. We are working on a studio overhaul, fixes to multilingual output, upgrades across the app, six brand new voices and native multilingual delivery for every voice across all 23 included languages.
This is a preview, not a launch announcement. None of the upcoming changes described here should be read as available now. For what has actually shipped, use the changelog. I want to explain the direction without asking you to plan work around a promise I haven't made.
What is coming next for Vois?
The work has a connected purpose: make the studio a better place to finish voice projects, and make the voice choices more useful across languages. I don't see those as separate ambitions. A voice library matters because of the projects you can complete with it.
The studio overhaul is coming. So are multilingual output fixes, upgrades across the app and the new voices. Alongside that, we are working on every voice getting native multilingual delivery across the included language set.
I'm keeping this note to those commitments. A preview can easily become a list of possibilities, and possibilities start sounding like promises once they're written beneath a product name. I would rather be clear about the work underway than give you a longer list to speculate about.
If you make recurring voiceovers, the direction should feel relevant to ordinary production, not just a special demonstration. Picking a narrator, keeping the delivery appropriate and finishing the piece are the moments I want this work to serve.
Why are we working on a studio overhaul?
I think a voice app should be judged by what happens after the first good take. You still have to make decisions about the script, the delivery and the relationship between sections. A studio should help you stay with those decisions instead of pulling your attention toward the tool itself.
That's why the studio overhaul matters to me. It is work on the place where a project comes together, not simply another reason to generate an isolated sample. The intended benefit is a better environment for doing the editorial work that remains after the voice speaks.
I'm not announcing a detailed interface or a set of new controls here. Those details aren't the promise of this note. The commitment is that we are working on the studio itself, with the finished project in mind.
You can read our existing multi-track timeline guide for the current production approach. The overhaul is upcoming work, separate from that guide. I don't want a description of what we're building to blur the distinction between the studio you can use and the studio we're working toward.
The voice library is getting the same treatment. Every voice carries its own illustrated avatar instead of a letter in a circle, so a cast is something you recognise at a glance. Filters sit where you look for them: search, category, language, and a switch for male and female voices. Language support reaches the built-in voices too, so picking a language is part of choosing a voice rather than a separate setting somewhere else.
There is more in this pass than I want to list here, and I would rather you found the rest in the release.
What do the multilingual output fixes aim to change?
We are working on fixes to multilingual output because the language version is the work, not an accessory to it. If you're producing narration in another language, the result deserves to be judged on that language's terms rather than treated as a variation on an English sample.
My aim is for that work to support more confidence in the output. I'm not going to turn that into a claim that every pronunciation or every passage will be perfect. A fix isn't a reason to stop reviewing the actual narration you intend to publish.
The useful test for you will still be your script: your names, your sentence structure and the delivery your listener expects. A polished demonstration can't answer all of those questions. When the changes become available, they should be assessed against real material rather than the excitement of an announcement.
These fixes are work in progress. I also want to distinguish them from the broader voice-language change below. Improving multilingual output and expanding native multilingual delivery across the voices belong together, but they aren't interchangeable descriptions of the same promise.
What will native multilingual delivery mean for every voice?
Every voice is getting native multilingual delivery across all 23 included languages. That is the direction of the upcoming release, not a statement about availability today.
The reason I care about this is casting. You might choose a voice because it suits a patient explanation or carries a story with the right restraint. I want the included languages to become part of that voice's usefulness, rather than forcing you to think of the voice and the language as disconnected choices.
That doesn't mean an identical performance in every language. Natural phrasing shouldn't be reduced to copying the rhythm of a source recording. A translated sentence can need a different shape to sound right, even when the voice is serving the same role in the project.
Keep that distinction in mind when reading this preview. Native multilingual delivery is the upcoming capability. Whether a particular read suits your audience remains an editorial judgment. The multilingual feature page is the place to check the current product offering; this founder note is describing where the included voice experience is heading.
Why are six new voices part of the work?
Six brand new voices are coming because casting is a real creative decision. Sometimes the script is fine and the delivery settings aren't the problem. The narrator simply isn't the right fit for the piece.
A wider choice is useful when it gives you a different way to approach the material. I don't want to describe these voices with invented personalities or announce roles before I'm ready to show the actual reads. The useful thing to hear will be how they handle a complete passage, not a label attached to a sample.
For you, the opportunity will be to audition the new voices against work you already understand. A familiar script is a good reference because you know where it needs warmth, clarity or restraint. It gives you something more concrete than a first impression.
There is no reason to recast an established project simply because a new voice arrives. If your current choice is doing its job, keep that decision. New options should create room for better choices, not turn an approved production into unfinished business.
What do upgrades across the app mean for this preview?
Upgrades across the app are part of the work underway. I'm deliberately not translating that phrase into an unannounced feature list. It would be easy to make this section sound bigger by naming possibilities, but that would make the note less useful.
My view is that product quality has to extend beyond whichever capability gets the headline. The app is where you make repeated decisions, revisit material and bring a project to a stopping point. Work on that broader experience belongs beside voice work, not beneath it as an afterthought.
The honest boundary is that this preview describes direction. It does not specify every change or claim a measured improvement. I'll leave concrete shipped details to the release record rather than imply that broad language is a guarantee about an individual workflow.
If a current project depends on a particular behavior, make the decision using the app that's available. Upcoming work is not a substitute for a capability you need to rely on now. That's true even when the work is something I'm personally excited about.
When will the next release be ready?
I'm not promising a date. There is no release schedule to extract from this post, and I don't want “coming next” to become a deadline in someone else's production plan.
The note is meant to make the direction visible: a studio overhaul, multilingual output fixes, app-wide upgrades, new voices and native multilingual delivery across the included languages for every voice. It isn't a request to delay a project until an announcement arrives.
Keep using shipped behavior as the basis for your work. When you're deciding whether to start a project, choose tools and voices that meet the brief today. When upcoming changes become available, judge them by what they let you finish, not by how confidently I described the intention.
That's the standard I want to hold myself to as well. I can share what we're working on without pretending the work is already done. And I can be excited about the next release without borrowing certainty from a calendar.
Keep making things worth listening to. I'll keep working on the studio around them.
-- Praney