Music AI Sandbox is a suite of experimental AI tools designed to empower musicians, producers, and songwriters by sparking new creative possibilities and helping them explore unique musical ideas. Developed by Google DeepMind in collaboration with YouTube and other partners across the music industry, it leverages machine learning algorithms to generate music loops and compositions from text descriptions. The goal is to integrate AI into music workflows in a way that enhances human creativity rather than replacing it.
Key features include Create, which generates music samples from natural language descriptions; Extend, which helps develop existing pieces by creating stylistically appropriate continuations; and Edit, which allows for reshaping music clips by changing mood, genre or style. The platform also features Lyria 2 (and the newer Lyria 3), Google DeepMind's music generation model that delivers high-fidelity outputs and nuance. Lyria 3 can generate its own lyrics and gives users more control over elements of the song, like style, tempo, and vocals. All outputs from Lyria 2 and Lyria RealTime are watermarked using SynthID, ensuring transparency and traceability.
Music AI Sandbox is not currently publicly available but is being shared with musicians, producers, and songwriters through YouTube's Music AI Incubator. Interested musicians can join a waitlist to experiment with the latest features. While specific pricing details for broader commercial use are not available, Google offers various plans for AI music generation, including pay-as-you-go options and subscription-based models for professional and business use.
