Mureka - AI Song&Music Maker

I’ve been using this app all Year to make my music ideas come to life. I LOVE this app, however the update on 9/9/25 is the worst! I only make 80s rock/metal songs, and this update is screwing me over on creating songs with the established sound I created this year. I hope they reverse the update and make it back to how it was before.

You can’t search for a friend if they are in there. And exactly how do you sell it to the crowd? You generate just about any types of genres and pretty much make your own (can’t name your own though). I truly enjoy playing with the app. I lost my voice a few years back and this sings for me. I also have Marfans Syndrome and I LOVE playing around with musical instruments but there are days I can’t lift up a guitar or my didgeridoo, so this is actually very helpful in making music and a sound to what I describe. The app gives two choices to choose from. I wish they had an Ai video generator- actually thought that they did that along with your music, it was the whole point of why I got the app. But no search, no generated video with music, and one more thing - I thought you could drop loops into sections like a DJ idk it’s good though, I enjoy it anyway :)

Subscription model is broken and way confusing. Everyone I click on Pro features I’m promoted by a paywall and I have to bypass by paying for something to then be greeted by a “you’ve already paid for this”. Then I click “create” and get redirected back as a loop. Sometimes prompted by a smaller native confirmation dialog to pay for some pro credits. I’ve paid for as much as I can when it pops. It’s simply a poor UX with bugs. Text prompt is 50% accurate. It doesn’t work half the time and generates random contexts. Lyrics sometimes mispronounce words on an otherwise perfect song. That’s likely an Eleven Labs or OpenAI voice model issue is my guess. But I would like to be able to regenerate based on the same context, but fix little things like word pronunciation by re-promoting. “Extend” feature doesn’t work for me. Pay wall loops. Male voice options are very limited. I also get female 90% of the time even when asking for male vocals. My guess is your trained models are not large enough or the prompts you shape are biased. I cannot open Zip exports of Instrumentals and Stems. Appears Zip is corrupt. On occasion I get random lyrics generated after I apply them to a generated song. My guess is some variable is getting reset to null or empty string and it is taking some GenAI alternative. Please fix bugs.