Skip to content
AnyCapAnyCap Docs

Music Generation

Live catalog

Supported music models

Choose a model before you generate. Availability and modes are loaded from the current AnyCap catalog.

View all modelsInstall the CLI

Loading live models…

    Modes

    Section titled “Modes”
    Mode Use it for
    text-to-music Create an instrumental track or song from a prompt

    The selected model may expose a different subset of controls. Check the live catalog before relying on a prompt field or optional parameter.

    Terminal window
    anycap music models \
      | jq -r '.models[] | .model as $model | .operations[]?.modes[]?.mode | [$model, .] | @tsv'

    Generate music

    Section titled “Generate music”
    Terminal window
    anycap music models
    anycap music models <model-id> schema --mode text-to-music
    anycap music generate \
      --prompt "warm lo-fi study beat" \
      --model <model-id> \
      --tags "lofi,warm,study" \
      --instrumental \
      -o study-beat.mp3

    Use --title, --lyrics, and schema-advertised parameters only when the selected model supports them. A request can return multiple clips under outputs.

    Terminal window
    anycap music generate --prompt "..." --model <model-id> \
      | jq -r '.outputs[].local_path'

    Output metadata

    Section titled “Output metadata”

    Music generation can return one or more output clips, along with duration, byte size, usage, and diagnostic IDs. Inspect every outputs[] item rather than assuming a single file.

    For speech, dialogue, and audio scenes, see Audio generation.