Subscribe
Sign in
Revoice — Your voice, powered by AI
Revoice from Podcastle brings generative AI to audio creation, allowing you to create a digital copy (i.e. clone) of your own voice. Built with safety & security in mind, the tech ensures only your own voice (not anyone else's) can be replicated.
32
Replies
Congratulations Team Revoice! The launch of this ingenious product on Producthunt is really exciting. It’s thrilling to see AI being utilized for creative audio creation in such a unique and secure manner. One suggestion - Maybe in future updates, you could add a feature where users could slightly modulate or alter their digital voice clone for different types of content or emotions? It could add an interesting twist! Keep going, fantastic job!
@manmohit thank you for the kind words, and great suggestion! Our team is already on it ;)
Congrats on the launch of Revoice! Exciting stuff! With its AI-powered voice transformation, this cool new product has the potential to revolutionize how we communicate. I'm curious, who do you think would be the primary users of Revoice?
@lucas_pilzen the current version is English-only, but hopefully more languages to come soon!
Revoice's voice modulation features are intriguing! How does it handle different accents and languages for a truly diverse vocal range?
Revoice's creative potential is noteworthy! How does it handle voice modulation in real-time applications like livestreaming or video conferencing?
@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!
This sounds absolutely revolutionary! How accurate are the digital representations compared to the original voice recordings? Is there a limit to the amount of content one can produce? Is there a way to add emotions to the voice?
@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!
Congratulations on the launch! This is really amazing, and I appreciate the emphasis on reducing the risk of unauthorized voice cloning. Frankly, just developing capabilities to limit unauthorized voice cloning may become an entire industry of its own. Nicely done and I'm excited to see where it goes from here.
@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.
Big congrats on the launch, while I have seen similar products, this one stands out with its incredible design and beautiful aesthetics from the website to the product hunt launch photos! design can definitely make a difference!
Good luck! 🚀
Congrats on the launch, an exciting concept. And kudos on the nice presentation visuals ❤️
Love the idea! Curious to learn what the tech does bts to avoid cloning voices other than your own.
@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.
Podcastle