Google’s Flow Music adds song generation, editing and stem splitting to Spaces, its ‘vibe coding’ tool
Google has upgraded Spaces, the "vibe coding" tool within its Flow Music platform, adding the ability to generate and edit full songs, split recordings into stems, and create images, lyrics and videos directly within the tool. The update, announced on X on 24 July, marks Spaces' evolution from a system for building custom instruments and effects into a more complete music production environment, bringing Google's offering closer in line with rival AI music platforms.
New song features let users create covers, replace sections and extend existing tracks. Flow Music is powered by Google DeepMind's Lyria 3 Pro model alongside Gemini, Veo and Nano Banana for chat, video and image generation, with all output watermarked using SynthID. The platform originated as the open-source Riffusion project before becoming ProducerAI and then Flow Music following Google's acquisition; in May, distributor Believe partnered with Google to offer the tool to its artists. The additions bring Flow Music's capabilities closer to those of Suno, which launched its Suno Studio DAW in September 2025, and Udio, whose Sessions editor debuted in June 2025.
- Google adds song generation, editing and stem splitting to Flow Music's Spaces tool
- Update announced 24 July, builds on earlier ProducerAI/Riffusion origins
- Move puts Flow Music in closer competition with Suno and Udio