Generate a song using preset auto-style prompts.
This node uses pre-trained style tokens to condition generation
without needing reference audio. Choose from genres like Pop, Rock,
Jazz, etc.
Build style description string from individual components.
This helper node makes it easy to construct a description string
that controls the musical style of generated songs.
Output format: "female, warm, pop, emotional, piano and drums, the bpm is 120"
FL Song Gen · 1 output
FL_SongGen_Generate
FL_SongGen_Generate
Generate a song from lyrics with optional text description conditioning.
This is the main generation node. Provide formatted lyrics and optionally
a style description to generate a complete song with vocals and instruments.
FL Song Gen · 1 input · 8 parameters · 3 outputs
FL_SongGen_LyricsFormatter
FL_SongGen_LyricsFormatter
Build properly formatted lyrics from song sections.
This helper node makes it easy to construct lyrics in the format
required by SongGeneration without knowing the exact syntax.
Format: Sections separated by " ; "
Lyrics within sections separated by "."
FL Song Gen · 1 output
FL_SongGen_ModelLoader
FL_SongGen_ModelLoader
Load SongGeneration model with variant selection and memory options.
This node loads the AI song generation model. Different variants offer
trade-offs between quality, speed, and VRAM requirements.
FL Song Gen · 1 output
FL_SongGen_StyleTransfer
FL_SongGen_StyleTransfer
Generate a song using reference audio for style transfer.
Provide a reference audio clip (ideally a chorus section) and the model
will generate new music matching that style with your lyrics.