Music and sound effects
Call list_audio_generation_models() first to inspect downloaded models.
Do not guess a model name or silently download a multi-GB model. Missing models
should be reported so the user can choose whether to install them in Lemonade.
Generating audio can load a multi-GB generator and evict the chat model from GPU memory. Only generate when the user explicitly requests audio, and expect the chat model to reload afterwards. Do not pre-generate assets.
For music, call generate_music with an ACE-Step model. Describe style,
character, instrumentation, mood and tempo in prompt. Leave lyrics blank for
instrumental background music; for vocals, use [verse], [chorus] etc. and
set vocal_language if known. Keep steps unset unless explicitly requested.
For sound effects, use generate_sound_effect with ThinkSound (or a supported
OpenMOSS SFX model). Describe the physical sound and acoustic environment. cfg
is a ThinkSound-specific knob and should be left unset otherwise.
Both tools write WAV files and return an absolute audio_path. Check
status == "success" and that path before reporting a completed asset.
If using Ren'Py, copy the generated files into game/audio/ through an
approved file operation, and reference the relative paths from Ren'Py scripts.
Do not fabricate success, silently substitute a model, or retry indefinitely.