FloatSong is an AI music generation platform that allows users to create songs using multiple music model families, including Suno, Flow Music, YuE2, and Mureka. The primary function of the tool is to enable users to input the same musical idea—whether as a text description, lyrics, or a combination—and generate separate takes across different models. Users can then compare the results and select a preferred version to develop further, including turning it into a music video. The platform is accessible via the web at floatsong.com. The tool offers a range of capabilities beyond basic song generation. Users can create music from text prompts or their own lyrics, generate lyrics for verses and choruses, and change a song’s style or arrangement, which may also alter melody and vocals. Additional features include extending a song with new sections, covering an existing song, and training a singing voice from recordings of the user’s own voice, which can then be applied to a recording with pitch controls. FloatSong also provides audio processing tools such as a vocal remover for separating vocals from music, a stem splitter for isolating individual instrument tracks, and a mastering option that processes audio on the user’s device. Music video creation is available, allowing users to upload a song, select a section, and generate an MP4 video; this feature requires an active subscription. Typical use cases involve describing a song idea, choosing a music model, listening to the generated tracks, and downloading a preferred take. Users can work in Simple mode with a short scene description, Custom mode for detailed lyrics and style, or Instrumental mode for backing tracks. Drafts are stored in the browser during exploration, while submitted songs appear in the user’s account library with status, playable results, and download options. The platform is designed for iterative creative workflows, such as generating music for a travel montage, product reveal, or closing scene, and requesting instrumental versions when dialogue needs space. Technical requirements include an active subscription for certain features, and some audio processing tasks are performed locally on the user’s device.