New Apps Songburst and CassetteAI Transform Text into Custom Music Clips
With the rise in popularity of Large Language Models (LLMs) and Generative AI tools like ChatGPT, developers have found new use cases to mold text in different ways, from writing emails to summarizing articles. Now, they have taken it a step further by creating apps that generate bits of music simply by typing in some words.
Developer Brett Bauman recently launched an app called Songburst on the App Store. The app is designed to be user-friendly and doesn’t require a steep learning curve. Users can generate a music clip by typing in a prompt like Calming piano music to listen to while studying or Funky beats for a podcast intro.
In case users can’t think of a prompt, the app offers a variety of prompts in different categories such as video, lo-fi, podcast, gaming, meditation, and sample. The music is generated through Leap, with Bauman building the app’s backend using Vercel. Currently, there is a limitation of generating 30-second clips, and the quality may vary. However, Bauman plans to increase the length of the generated music clips and improve the overall quality over time.
Songburst offers a free trial, but there is also a subscription option available at $9.99 per month or $79.99 per year. Subscribers receive 20 song credits per month and have the ability to download tracks in mp3 format. Additional credits can also be purchased in packs of 5 ($7.99), 10 ($11.99), or 20 ($15.99).
Bauman’s aim in building Songburst was to create a simple and mobile native text-to-music solution that doesn’t rely on spammy tactics to generate revenue. However, he is not alone in this endeavor. Akhil Tolani, the creator of apps like Rapchat, has launched CassetteAI, which is available on the web and App Store.
CassetteAI operates similarly to other text-to-music apps, where users can enter a prompt and receive a track in return. However, CassetteAI can generate a much longer sample, up to three minutes. Tolani explained that the app relies on a custom model based on a seq2seq hierarchical architecture, trained on a specialized dataset to generate copyright-free music.
The app also provides tools for users to create different versions of the generated tracks, allowing them to edit and mix the music to create new compositions. While these tools are basic for now, they provide a starting point for users’ creativity.
Previously operating on a waitlist basis, CassetteAI has now opened up to more users. Tolani is also introducing a Cassette AI Pro subscription priced at $4.99 per month, offering unlimited song generation and access to better quality AI models for improved results.
Tolani believes that Cassette AI stands out from other music generators like Mubert and Beatbot due to its ability to produce higher quality music with a quicker turnaround time. He aims to ensure that Cassette AI is viewed as a tool for music creation, rather than a replacement for human creators. Tolani considers AI to be a valuable tool that makes music production accessible to everyone, while still respecting the ethical boundaries of the music industry.
In addition to indie developers, major tech companies have also entered the field of text-to-music generation. Google introduced its MusicLM tool during the Google IO developer conference in May, while Meta open-sourced its AI-powered music generator called MusicGen in June.
Despite improvements in the quality of the generated tracks, concerns remain regarding the training data used by these models. OpenAI, for example, made its Jukebox model partially open-source and prohibited users from creating music for commercial use to avoid legal issues. However, there are also AI-forward musicians like Grimes, who invited fans to create songs using her voice and split royalties with her.
These new text-to-music generation tools are primarily focused on creators who need copyright-free music for their videos or podcasts. The developers hope that musicians will also take notice of these tools and incorporate them into their sample or song-making processes.
Overall, as the capabilities of AI continue to advance, the possibilities for text-to-music generation expand. Developers are constantly pushing the boundaries to provide accessible and user-friendly solutions for music creation while respecting the creative process and ethical considerations.