In recent years, Large Language Models (LLMs) and Generative AI tools like ChatGPT have gained immense popularity among developers. These tools have proven their usefulness in various applications, ranging from writing emails to summarizing articles. Now, developers have taken it a step further by creating an app that allows you to generate music just by typing a few words.
A Simple and User-Friendly Music Generator
Developed by Brett Bauman, the creator of PlayListAI (formerly known as LinupSupply), the new app called Songburst has made its debut on the App Store. The app boasts a user-friendly interface that requires minimal effort to operate. All you need to do is enter a prompt such as “Calming piano music to listen to while studying” or “Funky beats for a podcast intro,” and the app will instantly generate a music clip for you.
Categories and Prompts
If you’re feeling uninspired and can’t come up with a prompt, Songburst has got you covered. The app offers a variety of prompt categories including video, lo-fi, podcast, gaming, meditation, and sample.
Backend and Music Generation
Bauman revealed that he built the app’s backend using Vercel, and the music generation process is powered by Leap. At present, there is a limitation of generating 30-second music clips, and the quality of some outputs may not be optimal. However, Bauman intends to address these limitations over time by increasing the length of the generated music clips and improving their quality.
Free to Try with Subscription Options
Songburst is available as a free trial, allowing users to experience its capabilities firsthand. However, the app also offers a subscription plan priced at $9.99 per month or $79.99 per year. Subscribers gain access to 20 song credits per month and have the ability to download tracks in mp3 format. Additionally, users can purchase extra credits in packs of 5 ($7.99), 10 ($11.99), or 20 ($15.99).
Avoiding Spammy Tactics and Creating Native Solutions
Bauman designed and developed Songburst to cater to users who seek simple and mobile-native text-to-music solutions without relying on spammy tactics to generate subscription revenue.
Competition and Alternatives
While Bauman is not alone in his pursuit of a text-to-music app, one notable competitor is Akhil Tolani, the creator of popular music collaboration app Rapchat. Tolani has recently launched CassetteAI, available on the web and App Store. Similar to other apps, CassetteAI generates music based on user prompts; however, it distinguishes itself by producing samples up to three minutes in length. The app operates on a custom model built on seq2seq hierarchal architecture, trained on a specialized data set to ensure copyright-free music generation.
Editing and Mixing Tracks
CassetteAI offers an interface that enables users to create different versions of generated tracks and provides basic editing and mixing tools to facilitate the creation of new compositions. While the editing capabilities are relatively simple, they still offer the potential for users to experiment and personalize their tracks.
Subscription and Quality Improvements
Tolani plans to introduce a Cassette AI pro subscription priced at $4.99 per month, granting users unlimited song generation and access to superior AI models for enhanced music production. He emphasizes that Cassette AI outperforms other music generators like Mubert and Beatbot with better music quality and quicker turnaround times. Tolani aims to uphold ethical boundaries within the music industry and position AI as a tool to assist music creators rather than replace them.
Targeting Creators and Fostering Collaboration
These text-to-music tools primarily target creators who seek copyright-free music for their videos or podcasts. The developers hope that musicians will adopt and incorporate their tools into the sample or song-making process, fostering collaboration and innovation within the music community.
Major Tech Companies Join the Movement
Aside from independent developers, major tech companies are also exploring the text-to-music generation trend. Google unveiled its MusicLM tool during the Google IO developer conference in May, while Meta open-sourced its AI-powered music generator named MusicGen in June.
The Challenge of Training Data and Legal Considerations
Despite improvements in the quality of generated tracks, concerns persist regarding the training data used by these models. OpenAI has taken steps to avoid legal complications by partially open-sourcing its Jukebox model and prohibiting its users from creating music for commercial purposes. Notably, some musicians, like Grimes, have taken an AI-forward approach by inviting fans to collaborate in creating songs using her voice and sharing royalties.