With the rise in popularity of Large Language Models (LLMs) and Generative AI tools like ChatGPT Developers have found use cases for shaping text in a variety of ways for use cases from writing emails to summarize the articles. Now, they are looking to help you create pieces of music by just typing a few words.
Brett Bauman, the developer of PlayListAI (formerly LinupSupply), launched a new app called Songburst on the App Store this week. The app does not have a steep learning curve. You just need to type a prompt like “Calming piano music to listen to while studying” or “Funky beats for podcast intro” to let the app generate a music clip.
If you can’t think of a prompt the app has prompts in various categories including video, lo-fi, podcast, game, meditation, and sample.
Image Credits: Songburst
Bauman told TechCrunch that he built the app’s backend using Vercel and the music was produced through Leap. Currently, there is a generation limit of 30 seconds and some outputs may be of poor quality. Bauman said that over time he will look at increasing the length of the generated music clip and improving the quality.
Songburst is free to try but it offers a subscription for $9.99 per month or $79.99 per year. The subscription gives you 20 song credits per month and the ability to download tracks in mp3 format. Users can also purchase additional credits in packs of 5 ($7.99), 10 ($11.99) or 20 ($15.99).
Bauman said he built the app because there are few simple and mobile native text-to-music solutions around that don’t use spammy tactics to get subscription money.
He’s not alone in trying to create a smooth text-to-music app, though. Akhil Tolani, who makes apps like the music collaboration app Rapchat, launched CassetteAI, which is available on the web and the App Store both.
At the input level, CassetteAI works similarly to other apps. You type in a prompt for music and it spits out a track. However, it can generate a sample up to three minutes long. The app creator says this is because the app works with a custom model based on the seq2seq hierarchal architecture and it is trained on a special data set to generate copyright-free music.
Image Credits: Boxes AI
The tool also provides an interface for users to create different versions of the generated tracks and edit and mix them to create a new track. These tools are pretty basic, so don’t expect to create a multilayered master track from them yet.
Cassette AI interface for mixing tracks Image Credits: AI cassette
Tolani said the tool has been operating on a waitlist basis until now, but he is opening it up to more people now. He told TechCrunch that he also experienced a subscription to Cassette AI pro priced at $ 4.99 per month, which will give users access to unlimited song generation, access to better quality models of AI for improved song generation.
The developer mentions that Cassette AI is better than other music generators like Mubert and Beatbot because it generates better quality music with a faster turnaround time. He added that with Cassette AI, he wants to respect the ethical boundaries of the music industry.
“We want people to see AI as a tool for making music, not a replacement for creators: calculators don’t replace mathematicians, they make it easier to calculate things. We want to make it music production is accessible to everyone for any use case,” he said.
These tools mainly target creators, who can use copyright-free music in their videos or podcasts. The developers also hope that musicians will notice their tools and integrate them into their sample or song creation process.
Besides indie developers, big tech companies are also taking a crack at the text-to-music generation problem. Google made the MusicLM tool public during the Google IO developer conference in May. In June, Meta open-sourced its own AI-powered music generator called MusicGen.
While the models are improving when it comes to the quality of the generated tracks, there are concerns about the training data they use to create music. To avoid legal troubles, OpenAI made its Jukebox model partially open-sourced and prohibited users from creating music for commercial use cases. Then there are some AI-forward musicians like Grimes, who invited fans to create songs with her voice and split the royalties with her in April.
