Other Software Maghni AI by Crescendia

[lIlI]

Vocatone and Misbah Studios are developing a new AI-based singing synthesis engine, intended for crowdfunding.

2025 Update: Development has been passed to Crescendia

​

About the project

VocaTone Studio (formally Vocatone) previously created VOCALOIDs Oliver and Yohioloid in collaboration with PowerFX
Misbah Studios describe themselves as: 'A company focused on cultural preservation via technology'

Q&As

Update Timeline

Dec 25 2020
- Vocatone announces the project and a Kickstarter for 7th of Jan

Jan 2nd 2021
- Vocatone announces that the Kickstarter will be delayed by six months

Jan 4th 2021
- Start of production

Feb 1st 2021
- Monthly progress report

Mar 1st 2021
- Monthly progress report

April 3rd 2021
- Monthly progress report
- 6 launch libraries and 7 launch languages announced

May 4th 2021
- Monthly progress report

June 4th 2021
- Aurum and Audine redesigns announced

June 6th 2021
- Monthly progress report

July 1st 2021
- Monthly progress report

August 2nd 2021
- Monthly progress report

September 9th 2021
- Monthly progress report and first sample released

October 2nd 2021
- Monthly progress report

November 16th 2021
- Monthly progress report

December 15th 2021
- Monthly progress report

December 22nd 2021
- Vocatone renamed to VocaTone Studio, Aurum and Audine's redesigns were previewed, and VocaTone's official website was launched.

January 16th 2022
- Progress update blog

February 11th 2022
- Progress update blog

March 7th 2022
- Progress update blog

May 22nd 2022
- Progress update blog

August 2nd 2022
- Progress update blog

March 5th 2023
- Twitter update

May 1st 2023
- First prototype complete

September 21st 2023
- Progress update blog

December 21st 2023
- IndieGogo campaign announced for launch on the 26th

28th September 2024
- Progress update blog

29th September 2024
- Kotarou announced for Maghni AI

11th December 2024
- Update thread from Cinna

23rd January 2025
- Update blog

25th January 2025
- UI design update

21st February 2025
- Update blog

1st March 2025
- Kotarou's design revealed

15th March 2025
- VRILLIANT announces RVBY for Maghni AI

11th August 2025
- Update blog

Discord clarification

This indicates that Misbah Studios and Vocatone have left the project as full time developers, meaning the departure of Natalia Shmueli, Barham Karim, and Austin Grissom. *edit* A backer and forum user alleges Natalia is still part of the project here. No official staff list has been released as of 15th October, 2025.

13th July 2026
- Crescendia posts Maghni Flow (non-AI timing model) to Github

15th July 2026
- Update blog from Cinna

17th July 2026
- VocaTone posts an update regarding shipments and refunds.

21st July 2026
- Update and Q&A from Crescendia

30th July 2026
- Update blog

Update Videos

SAMPLES

Little Dark Age
Want You Gone
Backdoor

DEMOS

Oliver:
'Right Where You Are' (English)
Raw engine output comparison

'Kodoku na Heaven' (Japanese)

'amoeba' (English)
Raw engine output comparison

'Dinky Dink (SAWTOWNE REMIX)' (English)
Raw engine output comparison

Audine:
'Answer Book' (English)

allegations

12th July 2026
- VocaTart (Crescendia developer) replies to an open letter by TakoyakiDonuts

14th July 2026
- Natalia Shmueli (Former CEO of Misbah Studios) posts a statement regarding her experiences working on Maghni AI

14th July 2026
- Suzie (artist and Maghni AI collaborator) posts a tweet thread regarding her experiences working on Maghni AI

16th July 2026
- Austin Grissom (Former CEO of VocaTone) posts a statement regarding his experiences working on Maghni AI

16th July 2026
- Natalia Shmueli releases a statement in response to Austin Grissom

17th July 2026
- Cocoa (Crescendia developer) responds to allegations in a Twitter thread

17th July 2026
- Suzie responds to Cocoa's Twitter thread

18th July 2026
- Natalia Shmueli responds to VocaTone's statement regarding refunds

19th July 2026
- Crossyloid (Former illustrator for Maghni AI) responds to Natalia's statement

21st July 2026
- The current Maghni AI team posts their timeline of events and responds to allegations

IndieGogo Campaign​

 

[xuu]

Aurum does indeed seem to be Ausgris. Very interesting, though as assumed they are making their own software rather than working with a pre-existing one as before. Will be reserving judgement until I hear a sample / proof of concept / anything since this is very, very ambitious.

 

[lIlI]

Random thoughts:
It's going to be controversial, but I prefer the name Aurum to Ausgris - it's more pleasant sounding and easier to pronounce. Aurum reminds me of gold and auroras, whereas Ausgris reminded me of Australia (cool but random) and gristle (bleh). If Aurum is not Ausgris: sorry for just roasting your name, Ausgris.

I fear they may be overreaching with the amount languages, unless the technology already exists with the AI they're using. I hope it's a success and we at least get a few new ones we haven't had before.

Hopefully they can give us some kind of tech demo before starting the campaign. If that's not possible, hearing how the voices sound in UTAU would be nice.

Finally:

please allow me a small moment of screaming

AAAAAAA ANOTHER SOPRANO WHYYYYYY most fans have been begging for a new powerful deep voice for so long but every time the feminine vocals are high pitched even from Western companies agehsoaghpjd i love girly girls but PLEASE give me a voice that can kick my ass SOMEBODY

 

[Twillby]

I missed the AI part of the name for a minute oops. I was thinking, “Well I mean I’ve seen UTAU banks for a lot of these languages” but if it’s AI then... yeah it really does seem a bit too ambitious for a new synth. I still kind of want to give them the benefit of the doubt for now but tbh at the same time I’m not feeling confident enough to consider backing it.

 

[Blue Of Mind]

I agree that the amount of languages listed feels too ambitious for a new synth, but my lizard brain is still getting excited over the idea of German and Russian support (I've given up on Vocaloid adding more languages at this point).

 

[___]

"A new singing synthesis engine, all ready..."

Great! Then show us : ))))))

 

[lIlI]

Yo Audine has horns? I thought they were headphones. Neat!

Vocatone just posted a large update on their Facebook, mostly going over what we already know. I've summarised some new details:

1. They have been working on Maghni AI for the past three years
2. Like other AI programs, it uses machine-learning, but retains editing abilities for the user
3. Aurum and Audine were recorded by professional vocalists familiar with synthesis engines
4. They aim to make the program easy to use and affordable
5. Merch will be among the reward tiers, and preorders
6. Kickstarter drops Janurary 7th, Q&A on December 27th
7. They have one more surprise coming (I'm betting it's a Sonika update!)

 

[VocAddict]

I'm not sure how I feel about this whole thing. I'm always happy to see new vocal synth projects being worked on but it seems a bit overly ambitious? From the FB post, they say that they have been working on it for 3 years and they have already have full support for quite a range of languages but despite all of that, they could not share a single vocal sample? Even if it's not for the two vocals they showed, at least something developmental to give an idea of what the synthesised vocals can sound like?

I just feel uncomfortable with the fact that a company is posting an announcement, especially when it involves money on top of the fact that this has been a work in the making for years and it was just so poorly done. The social media pages for the company they're working with (Misbah Studios) were made newly made/revamped in the past two weeks so there's no information concerning the company and don't even let me start on their website

Sorry if I'm sounding pretty pessimistic but my hopes aren't that high at the moment. Hopefully I'll have a change of heart on the 27th.

Also, can we please please please stop getting soprano voices like smh we have enough of those already.

 

[___]

I'm not sure how I feel about this whole thing. I'm always happy to see new vocal synth projects being worked on but it seems a bit overly ambitious? From the FB post, they say that they have been working on it for 3 years and they have already have full support for quite a range of languages but despite all of that, they could not share a single vocal sample? Even if it's not for the two vocals they showed, at least something developmental to give an idea of what the synthesised vocals can sound like?

I just feel uncomfortable with the fact that a company is posting an announcement, especially when it involves money on top of the fact that this has been a work in the making for years and it was just so poorly done. The social media pages for the company they're working with (Misbah Studios) were made newly made/revamped in the past two weeks so there's no information concerning the company and don't even let me start on their website

Sorry if I'm sounding pretty pessimistic but my hopes aren't that high at the moment. Hopefully I'll have a change of heart on the 27th.

Also, can we please please please stop getting soprano voices like smh we have enough of those already.

Credit to @TheStarPalace for initially sharing the info but also.....a new job opening was posteted by Vocatone employee regarding new vocal synth MONTH AGO. The wording ( such as: "I'm currently looking for someone well versed in C++ and/or Python for building a vocal synthesis DAW." and "We plan to use Kfr or JUCE’s framework" ) makes it sound like the work hasn't even started on the engine itself.

I have hard time believing Vocatone.

 

[VocAddict]

Credit to @TheStarPalace for initially sharing the info but also.....a new job opening was posteted by Vocatone employee regarding new vocal synth MONTH AGO. The wording ( such as: "I'm currently looking for someone well versed in C++ and/or Python for building a vocal synthesis DAW." and "We plan to use Kfr or JUCE’s framework" ) makes it sound like the work hasn't even started on the engine itself.

I have hard time believing Vocatone.

I forgot about that listing, my hopes in this project is dropping by the second

 

[___]

Also sorry to be bugging again but AI voicebanks haven't gained that big of a relevancy til 2020, especially in the west and I have hard time believing Vocatone came up with this idea of AI synth 3 years ago.

edit: granted it could've been sample based engine but if they were genuinely working on it I think we'd be hearing samples of the sample-based vbs rn and be promised AI support down the line.

 

[TheStarPalace]

This announcement felt very rushed. First they said it would be announced in 2021 and then pushed it out for Christmas. I honestly would have preferred that they waited until there was a sample or something substantial.

 

[cafenurse]

I feel like such a big debby downer because a new voice synth should be EXCITING especially considering it's by a western company and we don't really see new synths from the west much now but....I'm just not that excited. I was really hoping the new announcement would be a Vocaloid or a Synth V because those are synths I'm familiar with using and not to be an old fogie, but I'm kind of at the point where I know how to use the synths I like and I'm much more inclined to stick with them then try new things. Voice synths are expensive and can be daunting to learn and it just makes more sense for me to stick with the few I'm familiar with then constantly be trying out new ones.
I guess it also doesn't help how bare bones this announcement is. We have a new voice synth engine coming out! Here's pictures of some of our vocal characters. Please fund it! Like...if this is supposedly finished...do we not get even a few seconds of a sample? I want to know what their voices sound like at least a little bit. A screenshot of the UI? It's hard to get pumped when I know nothing.
Obviously, my entire outlook on this may change when they start showing off demos and their VBs and what the voice synth looks like and how it functions and blah blah, but right now my mood just based on the announcement is. Oh. Ok.
I'm not like...mad about it, but I don't really see myself backing it. I wish them luck and I hope I get excited about it later when we know more.

 

[morrysillusion]

very much agree with everyone here- its cool to see what they have in mind but nothing presented makes me think it will happen easily or at all. its quite ambitious to claim all that langauge support and to not have any kind of proof of the synth doing its thing. even if it wasnt a AI bank, just as a normal sample bank wopuld be a good grasp on what we will get by the end. i would LOVE to see AI synth stuff with english as i have yet to hear it/hear much of it from the main industry of vocal synths. and here they claim this, but like. nothing to show? i get they will be asking for funding to get it going but 3 years of work with no prototype/alpha samples of give us an idea.... but they are fully ready to hype fans up with the character art? idk.

if theyre expecting hype for the characters from fans to support their funds then thats a worry to me in the sense that the tech should be the first thing we heard/get an idea of before throwing money at it. of course said tech may be early in development but something is better than nothing. right now there is nothing, other than some drawings.

ill cautiously watch from a distance but im not giving them my money until i know wtf theyre gonna be able to make. considering Vocatone popped into existence after years and does this? yeah im not...quick to trusting lol. i hope they will come through with something, because it IS great to see more new companies entering the ring of vocal synths, mostly when it will be aimed at other languages. so i dont wish them bad luck. but ya know...cautious.

 

[WyndReed]

I agree with the points everyone has brought up, but my question is how do you pronounce Maghni?? Is it mag + ni?

 

[GreenFantasy64]

Quick thoughts, the guy somewhat interest me, but there is something off with the girl design... The music and video is nice by the way.

 

[Kona]

I have conflicted feelings about this, but I’ll keep thinking positive and hopeful for now! I don’t think I’m too interested in getting into another synth and another learning curve again, but I wish the best for the project and want to see how it all unfolds.

 

[Prism]

Does anyone know of the old samples of agris?

 

[___]

Does anyone know of the old samples of agris?

Here!

 

[Prism]

Is it bad I hope this isn't the voice provider we get