{ "name": "Makima", "id": "makima", "description": "Makima's Japanese-accented English voice for TTS synthesis.", "language": "en", "accent": "ja", "sample_rate": 24000, "format": "pcm_f32", "model_backend": "qwen3", "reference_audio": "reference.wav", "notes": "Default voice for the Makima system. Reference audio should be a short (5-15s) clip of the target voice at 24kHz mono. Place the WAV file as reference.wav in this directory." }