MiniMax Music 1.5: Full Songs Driven by Structured Lyrics

This piece uses the text-to-music pipeline of MiniMax Music 1.5 to generate a complete song with vocals and accompaniment, demonstrating how to obtain a ready-to-use finished track from just two required fields: a "style prompt" and "structured lyrics." Input composition: the top-level prompt serves as the style prompt describing genre, mood, and instrumentation; base.lyrics arranges the lyrics in sections using [intro]/[verse]/[chorus]/[bridge]/[outro] structure tags to constrain the song form's progression; bitrate, sample_rate, and audio_format use the schema defaults (256000 bps, 44100 Hz, mp3), producing standard audio without additional configuration. This pipeline has no audio reference material field—all audio content is generated by the model in a single pass. Use cases: theme song drafts for short dramas and video soundtracks, rapid prototyping of virtual artist demos, and finished tracks for creators without arrangement experience who work from mood prompts. This version does not support instrumental-only mode or automatic lyric writing, the lyrics limit is about 600 characters, and results are fairly sensitive to prompt and lyric quality—so this piece focuses on how to write the structure tags. This piece was genuinely generated by the author's pipeline and is well suited to creating this kind of content with the MiniMax music model family.