@george_kalandadze the key differentiator with respect to security is that Revoice requires users to submit a LIVE voice recording of 70 pre-determined sentences to generate the model input. Then our team will manually listen to the voice recording samples for consistency and accuracy of the voice, before feeding it into the AI model. Whereas most other (less secure) voice cloning services will allow you to submit any pre-recorded sample, leaving the door open for submitting & cloning a voice other than your own.

@manmohit thank you for the kind words, and great suggestion! Our team is already on it ;)

@jasdeep_chhabra We think they're pretty darn accurate, but our AI is constantly getting better and more realistic. You can see a couple of "before & after" examples midway down this page: https://podcastle.ai/products/re... Our Pro plan limits creators to 1M characters (~20 hours) of Revoice content generated per month, but custom/enterprise plans can exceed that.

@dmitry_kopylovsky That's awesome Dmitry! But can you add emotions to the voice too? Lets say I want to say a sentence angrily and another dramatically, can I control that?

@jasdeep_chhabra not in this version, but we're working on it for the (hopefully) near future!

@ashna_s Thanks! Podcastle does put a premium on safety & security. Revoice requires users to submit a LIVE voice recording of 70 pre-determined sentences to generate the model input. Then our team will manually listen to the voice recording samples for consistency and accuracy of the voice, before feeding it into our AI model. This is contrary to most other (less secure) voice cloning services, that simply allow you to submit any pre-recorded sample, leaving the door open for misuse and cloning a voice other than your own. Revoice requires a bit more effort on the part of the user, but we believe this friction is critical to prevent misuse.

@dmitry_kopylovsky Agreed that's great, gating might create friction but bring valuable users who are inclined towards the service.

@andriyuh a key differentiator with respect to both quality & security is that Revoice requires users to submit a LIVE voice recording of 70 pre-determined sentences to generate the model input. Then our team will manually listen to the voice recording samples for consistency and accuracy of the voice, before feeding it into the AI model. Whereas most other (less secure) voice cloning services will allow you to submit any pre-recorded sample, leaving the door open for submitting & cloning a voice other than your own. All that said, we will continue to update our AI models to upgrade the product quality, including tone modulation for emotions. Stay tuned for more updates!

@lucas_pilzen the current version is English-only, but hopefully more languages to come soon!