AI Agent Hub
Back to skills
MiniMax Music and Lyrics Generation icon

MiniMax Music and Lyrics Generation

Design & Media Updated 2026.08.29

Paste the following prompt into your AI chat to install this skill:

Please install @user_c9d52b8e/minimax-music-v2 according to https://skillhub.cn/install/skillhub.md.

About this skill

Problem Solved

Many engineering workflows need to turn a prompt into lyrics first, then into playable audio. Calling the MiniMax Music Generation API directly can introduce friction around parameter combinations, model availability, audio formats, local playback dependencies, and plan permission boundaries. This skill organizes the common calls into two command paths: lyrics generation and music generation, reducing manual request assembly.

How It Works

Key capabilities include:
- Lyrics generation: provide a theme via prompt and choose write_full_song, write_lyrics_only, or optimize_lyrics with --mode for full songs, bare lyrics, or existing-lyrics refinement. Section tags such as [Intro], [Verse], [Chorus], and [Bridge] can express song structure.
- Music generation: describe style, mood, or scene; optionally attach --lyrics; generate instrumental audio with --instrumental, and control output through --format, --bitrate, and --sample-rate.
- Output handling: audio defaults to minimax-music-output/, with --output for custom paths; lyrics print to stdout by default or are saved when an output path is set; --play can play the result locally after generation.

Boundaries

Note that the Music Generation API requires a separate music Plan, while the Lyrics Generation API can use a Coding Plan key. The verified model is music-2.6. The API Key may come from config.json or the MINIMAX_API_KEY environment variable. pyglet is mainly used for playing mp3/wav, so a playback stack is generally unnecessary for lyrics-only generation.

Use Cases

  • Draft full singable lyrics for a short-video script, then generate an instrumental demo from the same description.
  • Refine existing lyrics with [Verse] and [Chorus] tags, then render a 256 kbps MP3.
  • Create an instrumental ambient bed for a prototype demo, export WAV, and audition it locally.
  • Produce candidate melodies and lyric versions with music-2.6 before a creative review.

Best For

  • Short-video editors who need to turn copy into lyric and melody demos quickly.
  • Frontend engineers or PMs adding an audition-ready ambient music bed to prototypes.
  • Backend engineers wrapping MiniMax audio APIs and needing stable lyric generation parameters.
  • Game tech designers producing instrumental event snippets and validating local formats.