backend/synthesis_catalog.py

OMNIVOICE_MODEL = "omnivoice"
MAGPIE_MODEL = "magpie"

LOCAL_BACKEND = "local"
MODAL_BACKEND = "modal"

SYNTHESIS_MODELS = [
    {
        "id": OMNIVOICE_MODEL,
        "name": "OmniVoice",
        "desc": "Presets, design prompts, and voice cloning",
    },
    {
        "id": MAGPIE_MODEL,
        "name": "Magpie TTS",
        "desc": "Multilingual speaker presets with text normalization",
    },
]

SYNTHESIS_BACKENDS = [
    {
        "id": LOCAL_BACKEND,
        "name": "HF/local",
        "desc": "Run synthesis inside this Space runtime",
    },
    {
        "id": MODAL_BACKEND,
        "name": "Modal",
        "desc": "Offload synthesis to a deployed Modal worker",
    },
]

MAGPIE_SPEAKERS = [
    {"value": "Sofia", "label": "Sofia"},
    {"value": "Aria", "label": "Aria"},
    {"value": "Jason", "label": "Jason"},
    {"value": "Leo", "label": "Leo"},
    {"value": "John", "label": "John"},
]

MAGPIE_LANGUAGES = [
    {"value": "en", "label": "English"},
    {"value": "de", "label": "German"},
    {"value": "es", "label": "Spanish"},
    {"value": "fr", "label": "French"},
    {"value": "it", "label": "Italian"},
    {"value": "vi", "label": "Vietnamese"},
    {"value": "zh", "label": "Mandarin Chinese"},
    {"value": "hi", "label": "Hindi"},
    {"value": "ja", "label": "Japanese"},
]

MAGPIE_OPTIONS = {
    "speakers": MAGPIE_SPEAKERS,
    "languages": MAGPIE_LANGUAGES,
}