Dev Diary: Cantai 100% Local Rendering & Ethical AI for Dorico (April 30 Launch)

Dorico
1

Hi everyone,

As we approach the April 30 launch for Cantai for Dorico, we wanted to share a major technical update regarding Cantai.

Moving to 100% Local Rendering

To provide the most robust and private workflow possible, we have moved all of Cantai—including the phonemization—to your local hardware. Everything is local. Your scores, lyrics, and audio data never leave your machine. To ensure the best performance regardless of your setup, we’ve implemented a hybrid hardware hierarchy:

  1. Local GPU: Utilizes your graphics card for near-instantaneous synthesis.

  2. Local CPU: For machines without a dedicated GPU, the system runs entirely on your processor.

  3. Cloud Failover: We maintain an AWS path only as an emergency backup for legacy hardware.

Deep Integration for Dorico

We are building Cantai to feel like a native part of the Dorico environment. Our focus is on ensuring the engine responds seamlessly to your notation, so you can focus on composing rather than technical fiddling.

Ethical AI & Opt-in Stats

While Cantai is local and private, we are introducing an opt-in feature for anonymous usage statistics. We are committed to fairly compensating our singers; these stats allow us to track voice profile usage so we can distribute royalty cuts to those artists every quarter. You keep your music 100% private; they get paid for their work.

Try it Now in MuseScore Studio

If you want to find out what it will be like using Cantai in Dorico, you can try it right now in MuseScore Studio. It has been live there since December 1st and everything has been going fantastic. While MuseScore Studio itself is free, you can sign up for a month of Cantai for $14.99 to get a “hands-on” preview. It isn’t just about how it sounds—it’s about how it feels to use it. It is even better than the concept, and the responsiveness is amazing!

Pre-Order Update

Support for our $150 “Early Adopter” pre-order has been incredible. We are keeping this 50% discount active (MSRP $299) as we finalize the local implementation for the April 30 release.

Pre-order here: https://cantai.app/preorder_dorico_sibelius

We’ll be hanging out in this thread to chat about hardware specs, the integration process, or the new singers joining our roster soon!

Best,
Richard deCosta
Composer, Founder
Cantai Intelligent Singing Synthesizer / Turing Opera Workshop

2

Any possibility of having weekly updates on what is being accomplished with Cantai for Dorico? It will soon be 3 weeks since the last software update and, as I and others have shared, there are still many issues to be resolved. Since there is so much reliance on customer reporting of issues, it would seem appropriate to keep us informed a bit about the progress periodically.

I know that I have personally submitted several issues and received lots of gratitude for the submissions but no real feedback for most of them. Just a bit of communication about what is being accomplished would be nice for those of us who really, really want Cantai to succeed. Thanks!

3

Aren’t you expecting too much from them?

If they have to release more frequent, that involves more testing, more communicating, more building of installers, etc., etc. That’s time that can’t be put in solving bugs or adding more features.

You aren’t expecting a new release from Dorico every few week too (are you?)
And the Dorico devs aren’t announcing future features too.

4

I think satisfying (or attracting) Sibelius customers at present will give Cantai a stronger financial foundation to make improvements for all platforms.

5

I need to clarify. By weekly updates, I meant progress updates, not product updates. Just communication. And, if that is no more than “working on Sibelius version at the moment”, then, that would be perfectly fine. It’s the lack of communication given such rich communication after the Dorico compatible launch to nearly nothing…

6

I agree with you.