Cantai isn’t just another sound library—it’s a breakthrough in vocal synthesis. Powered by advanced machine learning, Cantai captures the subtlety of real human singing, from delicate dynamics to expressive phrasing. Whether you’re writing an intimate solo aria, a soaring choral passage, or a dramatic operatic scene, Cantai delivers voices that sound authentic, emotional, and true to your score. Note: Cantai's voice quality is continually improving, and the refined production versions featured in our online demos will be available soon.

Previews

What’s New?

1.0.0

Cantai initial release

Description

Cantai isn’t just another sound library—it’s a breakthrough in vocal synthesis. Powered by advanced machine learning and fully integrated into MuseScore Studio, Cantai not only captures the subtlety of real human singing, from delicate dynamics to expressive phrasing, but also sings your lyrics with clarity and emotion. Whether you’re writing an intimate solo aria, a soaring choral passage, or a dramatic operatic scene, Cantai delivers voices that sound authentic, expressive, and true to your score. Cantai is available through subscription and includes Solo Soprano, Alto, Tenor, and Bass voices, as well as full choir ensembles. Rendering is managed intelligently to ensure optimal performance and system availability. Usage Note: Cantai includes up to 180 minutes of rendering time per week, which resets every Monday. Because Cantai renders automatically as you write, your minutes can be used up quickly. Previously rendered sections will still play back even after the limit is reached. To conserve minutes, switch to manual rendering using the steps here: https://handbook.musescore.org/sound-and-playback/musesounds/online-sounds

Voice modeling library integrated into notation
English and Latin lyric support with more languages supported soon
Seamless integrated with MuseScore Studio and Muse Hub
Realistic phrasing and dynamic expression
Natural breath handling and syllable shaping
Includes Choir and Solo soprano, alto, tenor and bass

1.0.0

It's amazing how tools are improving every single day. I've been looking for a tool like this that's compatible with Musescore and I don´t have to export xml, or cut parts, etc... It's kind of dissapointed the only laguages available are english and latín. My native language is spanish and also Irecently write something in italian and I wanted to give it a shot, I gave me an idea but of course with english accent. Also I try with some english songs and it was much better but still need to fix some bugs bc is not recognizing some words... Also, I couldn't export an mp3 file, to show a sample to someone :( I think is promising and definetely I'll keep an eye on updates, but at the moment I don't really think it is worth to pay a monthly subscription at the moment, but if we improve the languages most likely I'll be there :)

repentant-toad987

At the time of this software being released, I just so happened to have completed a choral symphony, and since I unfortunately don’t have singers at my immediate disposal (as with a lot of composers, I’m sure), it’s nice to get some immediate feedback regarding the vocal music composers write. While I’m sure that the voices will be improved upon in due time, my only main gripe is the monthly subscription; I had actually preordered a one-time paid lifetime subscription, so that I wouldn’t have to worry about putting aside money every month; this is quite detrimental, especially for the younger composers who have a specialty in choral music; other than this though, it has delivered quite beautifully for me and I can’t wait for what’s to come in the future!

MaestroHenry18

1.0.0

I’ve always wished for a product like this, so in a way it really caught my attention. Unfortunately, it doesn’t quite deliver. The biggest drawback is the monthly subscription. I might understand a one-time payment, but paying every month for something of this kind simply isn’t worth it. Another problem is the soprano range, which is surprisingly poor. And I can’t overlook the English accent present in all voices. These issues make the product feel unfinished, almost as if it wasn’t tested with real musicians in mind. If these shortcomings were addressed, I genuinely believe it could become one of the best-selling products on MuseScore. But if the developers don’t fix them, I’m certain a better competitor will take its place soon enough.

Kem_cz

1.0

It’s a very cool product that will be very useful for operatic or classical scores. Unfortunately, as a musical theatre writer, these voices aren’t helpful to me at all. This product will be a 10/10 if/when musical theatre style voices are added!

WoodsyWoods9804

1.0

Impressed by the skills putting this library together. Works very well for opera scores and the opera Cantai voices but less so for music theatre, musicals and pop (the soft and ordinary voices of the tenors and barritone are not great). Unfortunately the soprano is pitchy and strained in the upper range - Ab is not high enough for most opera scores and sopranos can hit this note without strain in the realworld. For sustained notes over multiple bars, the pitch dropped in the final bar before scooping back up to the correct note (for any of the tenors) and there are some mid-range notes that are also pitchy. Not enough choice for sopranos, mezzo and no alto or bass (hope these are coming). The online sounds have to be reloaded every time I open MuseScore - not sure about the cache notes in the manual - doesn’t seem to be the case for me as when I first open score that has Cantai already used, the default piano sound is active - using MacOS 26.1

JMJJ

1.0.0

I wholeheartedly agree that this product itself delivers exceptional results, resolving numerous issues with vocal-heavy sheet music that previously required human intervention. I am personally acquainted with several opera singers who work with this product. However, its limitations are significant. Firstly, it operates on a monthly subscription model, which creates considerable uncertainty for users. Most critically, it imposes a weekly quota of just 180 minutes. Regrettably, there is no transparency regarding how these 180 minutes are calculated, meaning users can easily exhaust their allowance when initially testing and adjusting the product. After weighing the audio quality against the cost-effectiveness, I ultimately believe the developers need to offer a more reasonable usage plan that benefits consumers. I fully acknowledge that MuseScore has consistently been a latecomer achieving technological progress, but careful consideration is required for specific market strategies. For this particular product, I personally feel that removing the quota restrictions would be the optimal solution.

Partner Response: Fixed! We’ve already increased the token limit by 10x — sorry about the initial quota being too low. It was a conservative first calculation on our side. And thanks for your kind words about the quality!
floatingpitchfork453
Works With
MuseScore Studio
Rating
33%
Version
1.0.0
Released
2025

You May Also Like