Hello,

I have a few points of criticism to make :

I was caught out during my first test. While listening to my initial recording, I heard some background noise on my voice take. In my haste, I generated a minimax expense of 100,000 credits, which I mistook for 10,000 credits. For this first point, it would probably be nice to have a solution, to put it in bold and probably to add a space for 100,000, and to indicate that it will cost 100,000 out of 120,000 credits.

The second point is to create a kind of "favorites" to save the settings and adjustments that can be made with existing voices, or a "laboratory mode" to fine-tune the settings, because you quickly become a sorcerer's apprentice, burning through credits at high speed.

The third point is that some voices, like "Chris" in French, have a Canadian accent when the audio is generated, whereas there is nothing to suggest this will happen. When you choose a cloned voice from the library, the parameters and the model type are not imported, so you are often disappointed because you expected something else from the demo. This leads to a lot of tinkering and wasted credits.

The fourth point concerns importing sounds to clone a voice. I find it too restrictive. I regret that the M4A format that my phone generates does not work. I would add a user verification workflow for the imported sound, or a pre-analysis to avoid importing a muffled, poor-quality, or noisy audio file.

The fifth point is that when I set the interface translation to French via Chrome, it creates errors for me.

This is somewhat the tool I was waiting for, but it still has room for improvement with these small issues, and the interface deserves to be reviewed based on all the remarks I've made.

Good luck to you.