Hello, everyone.
Before I begin, I would like to thank you for your continued patience and support as always. Since the previous update over three months ago, many things have happened and to start, I'd like to take a moment to apologize. It has been a little over six months since I have taken over this project and began overseeing its development. It deeply disappoints me that I was unable to keep my promise regarding steady updates due to some soon-to-be-disclosed roadblocks.
From myself and the rest of the Crescendia team, we sincerely apologize and hope to soon hammer out these speed bumps as soon as possible to ensure this doesn't occur again. We owe it to you, our supporters, and the several companies that have worked tirelessly to develop voice libraries for our vocal synthesis software to move back onto a consistent schedule.
The transition and internal restructuring of the project we mentioned towards the end of last year has ended up taking far longer than we had expected and hoped. What was initially assumed to be a short adjustment period has turned into a complete handover of the project. In short, nearly all of the individuals who previously led Maghni AI and its public commitments, ultimately overseeing the Indiegogo campaign, are no longer involved.
In their place stands a new, much smaller team of talented individuals—ourselves—who chose to remain after the prior management departed. We have been working to recover, stabilize, and rebuild what was left behind. While we are not the originators of many of the project's initial promises, we have chosen to do what we can to fulfill those promises wherever possible, with limited access to the original codebase and infrastructure.
With That Said, Exciting Progress: Mandarin XLS
Amidst all of these challenges, we are proud to share some real development progress. We have been working on one key promise, XLS, diligently; especially testing Mandarin XLS with vocals who have no Chinese data recorded or finalized, and it is moving forward at a strong pace.
Mandarin is a uniquely demanding language for synthesis due to its structure, and we have worked closely with native-speaking and fluent voicers to take steps towards building a foundation that feels natural, expressive, and accurate. As one of the original promises made by previous management, XLS is something we are working diligently to implement, and we are proud to finally unveil our first test results with you all. Please, have a listen to Mo li Hua performed by an older ver. of OLIVER AI, utilizing an early alpha build of our cross-lingual technology!
Backer Rewards
We would also like to directly address the subject of Indiegogo physical rewards, which many of you have understandably expressed concern over.
Although many members of our current development team had no hand in planning or managing the original Indiegogo campaign, we have still chosen to take on the responsibility of seeing them and their fulfillment through. Several months ago, we personally paid out of pocket to cover the remaining production costs and ensured that all reward items were handed off to VocaTone, who has been responsible for delivering orders since.
At present, all rewards within our power have been paid for, shipped to VT, and are there awaiting fulfillment. While this has required significant time and financial resources—resources that otherwise would have been dedicated to software development—we felt that it was critical that we tackle this issue as soon as possible to put an end to the extended delay our supporters have been facing.
For those still awaiting delivery, we ask for your understanding and continued patience as VocaTone completes distribution, and to please extend all inquiries regarding these items to VocaTone going forward.
Where We Are Now
Just to go over everything again, Maghni AI is now being led under a team of both previous and new developers under a new structure, and with new priorities. Our focus is not only on continuing software development, but also on recovering the final components of the project's proprietary code, which still partially remains in the possession of previous management. Reclaiming that final piece is essential to moving forward at full capacity.
We understand how long this process has taken, and we know it has not always been easy to follow or to wait through, but we remain committed to delivering the software you were promised and working towards getting back on track so that we can do so with more reliable transparency moving forward.
Thank you again for your trust, your patience, and your belief in this project. While we cannot guarantee an immediate return to regular monthly updates at this time due to the aforementioned hurdles in our team's path, we'll do our best to overcome and keep everyone posted with new information when it becomes available. Please look forward to it!
All the best,
— Cinna
Co-Director, Maghni AI