
Google DeepMind has revamped its AI music creation tools - MusicFX DJ, Music AI Sandbox and YouTube's Dream Track experiment - with new designs and features that make it easier than ever for users to generate professional-quality tracks. The updates are part of DeepMind's broader push to innovate in the music technology space.
MusicFX DJ, a collaborative project developed with the help of multi-GRAMMY winner Jacob Collier, allows users to generate music interactively using simple text prompts and intuitive controls.
Unlike traditional DJ tools that remix existing tracks, MusicFX DJ synthesizes entirely new sounds. With a new, reimagined interface, improved audio quality, and expanded control features, creators can shape music as it evolves—from guiding instrumentations to crafting nuanced breakdowns and bass drops—all in real-time.
The platform employs advanced generative models that adapt text prompts dynamically. Rather than relying on a fixed command, users mix and modify prompts as they play, giving them creative control over the output. This lets even beginners improvise live DJ sets or explore diverse genres, instruments, and rhythms. As Collier describes it, MusicFX DJ lets creators "craft real-time sonic putty that’s endlessly surprising."
DeepMind has also expanded its Music AI Sandbox, a suite of experimental tools initially previewed at I/O earlier this year. Though not yet publicly available, Music AI Sandbox integrates the latest generative technologies from DeepMind, including models that power MusicFX DJ. Designed for musicians and producers, the Sandbox introduces features like loop generation and a new multi-track view to help users organize and refine compositions. These updates promise to significantly streamline music production workflows by offering advanced tools for seamless sound transformation and track in-painting.

Alongside these updates, YouTube’s Dream Track experiment—another DeepMind initiative—is now providing creators in the United States the ability to generate instrumental soundtracks for Shorts, guided by powerful text-to-music models. Dream Track uses a novel reinforcement learning approach to enhance audio quality while staying responsive to user inputs.
With these innovations, Google DeepMind is continuing its mission to democratize music creation. The tools, while designed with professionals in mind, are accessible enough for anyone to experiment and create—regardless of their musical background. And with real-time streaming of production-quality audio and features like session sharing, DeepMind aims to bring the thrill of live music improvisation to a broader audience.
DeepMind's generative AI tools aren't just about making music easier to create—they're about reshaping how we think about music itself, bridging the gap between technical production and spontaneous creativity. As these tools continue to evolve, they might just help forge a new era where musical creation becomes a collaborative, interactive experience for all.