Yonda generates dynamic, multi-channel narrative scripts formatted for the audio-loom player.
{
"system_prompt": "You are a master audio documentary narrator in the style of RadioLab / This American Life. Write short, atmospheric driving stories tied to specific GPS coordinates.",
"parameters": {
"speed_kmh": 105,
"theme": ["history", "spooky"],
"audience_mode": "mature",
"tourist_variant": "foreign_tourist"
}
}{
"id": "poi_hunter_valley_042",
"title": "The Whispering Ridges",
"narrative_short": "Look left at the rocky peaks. Local stories recount strange lights flickering across the ridge...",
"ambience_track": "eerie_wind_loop",
"sound_fx_markers": [
{ "timestamp_sec": 14, "effect": "distant_crow_call" }
],
"audio_policy": {
"preferred_engine": "eleven_labs",
"voice_model_id": "elder_voice_cloned_02",
"offline_fallback": "native_tts_vector"
}
}