agents::VoiceConfig Struct Reference

Configuration for the VoicePipeline. More...

#include <stt_pipeline.h>

Public Attributes
std::string	agent_endpoint
	Endpoint URL for this agent.
std::string	stt_endpoint
	STT Endpoint URL.
std::string	tts_endpoint
	TTS Endpoint URL.
std::string	model_path
	Path to ggml model.
std::string	vad_model_path
	Path to ggml VAD model.
std::string	language
	STT language (e.g. "en", "fr", or "auto" for auto-detect).
int	n_threads = 4
	Number of threads for whisper.
bool	use_gpu = false
	Use gpu - keep false on Intel macOS.
double	requested_sample_rate = 16000.0
	audio sample rate
int	pa_frames_per_buffer = 512
	PortAudio frames per buffer.
int	device_index = -1
	default input device
float	max_buffer_sec = 30.0f
	cap PTT length
bool	print_events = true
	print events to console
float	eos_fade_in_ms = 200.0f
	fade-in duration to avoid clipping
float	eos_silence_threshold = 0.008f
	energy below this means silence
float	eos_silence_ms = 700.0f
	silence duration to trigger end

Detailed Description

Configuration for the VoicePipeline.