Ten months after v1, we're super excited to share Sonauto v2 (Beta) with the world. Our new site is beautiful and includes all of the features you'd expect from a full music platform -- following, comments, playlists, and groups (like subreddits for music). Our new model can beat the quality and diversity of paid models from much larger companies while remaining completely free and unlimited. Best of all, the model isn't done training yet and we're continuing to update it with significant improvments every few days. The best way to experience Sonauto is to hear it for yourself, so visit our site and listen through the Staff Picks, then try prompting your own song! -- https://sonauto.ai/ Currently all research/development/training of the model and most of the website (shoutout to our summer intern Lucas) was completed over the last few months by myself and my cofounder Hayden. A unique part of our process is constant user feedback. We've been very open with our community throughout experimentation/training and funneled their extensive feedback in order to make our v2 model as fun as possible. For example, we're super proud of the audio quality but especially the diversity of our model. One of the most constant complaints we heard about our largest competitor was that their model turns every prompt into generic modern pop music, so we made sure our model could produce diverse results across thousands of genres, decades, and styles. We'd love to answer your questions related to our dev process!

I've been having some trouble with songs with lyrics! Is it possible that the generation will get confused if I use made-up words or languages other than English? Other than that, I'm really enjoying the experience! The music generation is on point. Congrats on this launch!

These guys are the real deal! They built all of this from scratch, including the model, with a team of just 2. And, magically, after all that training, their model does feel state-of-the-art. Congrats guys!

Wow! This is unreal. The voices are the best AI-generated voices I've ever heard. You guys nailed the "diverse styles/genres" goal. I'm speechless...

Wow, this is the most exciting project I’ve came across on this platform. I just quickly tried out some of the lyrics I had saved in notes and it already made a banger. Will be trying this out more tonight. Does this work with any other languages (Spanish, Russian) ?

So far the fav product i have come across on ProductHunt , Just loved the concept on how Sonauto brings the creative soul inside you by giving your thoughts and expressions a wing ! Just created a song out of a short prompt : https://sonauto.ai/song/3cc5aecd... & i am speechless , If i could have rated it somewhere then a whole 10/10 from concept to execution to everything of sunauto ! Lovely Product !!!

This is so cool guys, well doooooneee I gave it a random topic, with random names I wanted it to mention in the lyric and the melody is awesome Good luck

What languages are available for the songs?

I happily recommend Sonauto v2 Beta! It is an amazing app that allows you to create full-fledged songs based on your prompts, melodies, or lyrics. It's perfect for those who want to quickly generate an original song, whether it's for a birthday or other event.

Sounds super cool, will definitely try, all the best for your launch!!

Sounds impressive but is it really better than the big name paid models?

Super awesome! This team is cracked out of their mind!

Really a Good Product . i am a musician , it really helps me a lot

im using suno, why should i try this

Do you have video tutorial available for this?

Congrats.. I like this .. got to try it ..

Cheers to the whole team for making it to Product Hunt! I hope the launch brings you not only visibility but also valuable connections and users who love what you’ve built.

wow I love using it! Great job, huge congrats!

its cool, damn.