Skip to main content

Crate speechcore

Crate speechcore 

Source
Expand description

Reusable speech-to-text runtime components for Rust applications.

speechcore provides audio capture, Silero VAD segmentation, backend selection, model provisioning, and transcript streaming without any UI, clipboard, tray, or application-specific enhancement code.

Push-to-talk applications should start with SpeechEngine. Applications that need full realtime state/control can use RealTimeTranscriber.

Re-exports§

pub use backend::traits::TranscriptionResult;
pub use backend::create_backend;
pub use backend::BackendCapabilities;
pub use backend::BackendConfig;
pub use backend::BackendType;
pub use backend::QuantizationLevel;
pub use backend::TranscriptionBackend;
pub use backend::TranscriptionError;
pub use config::AudioProcessorConfig;
pub use config::CT2Options;
pub use config::CommonTranscriptionOptions;
pub use config::DebugConfig;
pub use config::GeneralConfig;
pub use config::ManualModeConfig;
pub use config::MoonshineOptions;
pub use config::ParakeetOptions;
pub use config::PostProcessConfig;
pub use config::RealtimeModeConfig;
pub use config::SpeechConfig;
pub use config::VadConfigSerde;
pub use config::WhisperCppOptions;
pub use config::SAMPLE_RATE;
pub use feedback::FeedbackEvent;
pub use feedback::FeedbackSink;
pub use state::AudioVisualizationData;
pub use state::BackendStatus;
pub use state::BackendStatusState;
pub use state::ProcessingState;
pub use transcription_stats::TranscriptionStats;

Modules§

backend
Multi-backend transcription abstraction
config
feedback
post_processor
Text post-processing for transcription cleanup
prelude
state
transcription_stats