Song.do

Song.do is a death metal generator tool. A prompt to song generator with detailed genre, mood, vocal, instrument and tempo controls, plus a lyrics writer and a vocal remover. Three models, MP3 downloads, and a free Lite tier make it a low risk entry point for personalized songs and social clips.

Song.do is a browser based AI song generator that turns a short text prompt or a set of written lyrics into a complete track. The workspace on the homepage is a single form: a description field, a lyrics box with an optional AI Lyrics Generator, a style field with an Enhance helper, an Instrumental toggle, and a model picker. Push Generate and the service returns a song with vocals, melody and instrumental arrangement already combined, so nothing has to be recorded, sequenced or mixed by hand. The style controls are the substance of the product. Roughly 40 genres are listed, from Afrobeats and Bossa Nova to Trap and World Music, alongside mood tags such as Cinematic, Melancholic or Uplifting, about 20 vocal options (male, female, choir, falsetto, growling, spoken word, multilingual), a 30 instrument list, and tempo bands from very slow to very fast. Additional fields cover styles to exclude, title, song description and a custom duration option. Three models are offered: v5 Pro for studio grade output, v5 Muse for experimental results, and v5 Lite, which is free. Two adjacent tools round out the workflow. The AI Lyrics Generator drafts lyrics from a song idea, with more detail in the prompt producing better verses. The Vocal Remover is a separate utility for separating a vocal from an existing audio file. Finished songs land in My Songs, where they can be played in the browser, shared, or downloaded as MP3. The homepage gallery shows public results in several languages, including Arabic, Spanish, Indonesian and French, which says something about how the tool is actually being used: personal gifts, short social clips, and hobby projects rather than album production. That is the honest positioning. Song.do competes with other text to music services on speed and simplicity rather than on fine control. There are no stem exports, no MIDI, no arrangement timeline, and no versioning beyond regenerating. For a personalized birthday track, a TikTok backing loop, or a fast demo of a lyric idea, that tradeoff is reasonable. For anyone who needs editable audio, the ceiling arrives quickly.