Crate polyvoice 

polyvoice

Speaker diarization library for Rust — online (streaming) and offline (file-based), ONNX-powered, and ecosystem-agnostic.

Designed to be embedded into any Rust application that needs to answer the question “who spoke when?”.

Quick start

use polyvoice::{OfflineDiarizer, DiarizationConfig, DummyExtractor};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = DiarizationConfig::default();
    let diarizer = OfflineDiarizer::new(config);
    let extractor = DummyExtractor::new(256);

    // Load 16 kHz mono f32 samples (e.g. from an audio file decoder).
    let samples: Vec<f32> = vec![0.0; 16000 * 10];
    let result = diarizer.run(&samples, &extractor)?;

    for turn in &result.turns {
        println!("{}: {:.2}s - {:.2}s", turn.speaker, turn.time.start, turn.time.end);
    }
    Ok(())
}

Re-exports

pub use features::FbankConfig;

pub use features::FbankExtractor;

pub use cluster::SpeakerCluster;

pub use embedding::DummyExtractor;

pub use embedding::EmbeddingError;

pub use embedding::EmbeddingExtractor;

pub use offline::OfflineDiarizer;

pub use online::OnlineDiarizer;

pub use overlap::OverlapRegion;

pub use overlap::detect_overlaps;

pub use pipeline::Pipeline;

pub use pipeline::PipelineError;

pub use silero_vad::SileroVad;

pub use types::Confidence;

pub use types::DiarizationConfig;

pub use types::DiarizationResult;

pub use types::EmbeddingDim;

pub use types::SampleRate;

pub use types::Seconds;

pub use types::Segment;

pub use types::SpeakerId;

pub use types::SpeakerIdRemap;

pub use types::SpeakerTurn;

pub use types::TimeRange;

pub use types::WordAlignment;

pub use types::remap_segments;

pub use types::remap_turns;

pub use vad::EnergyVad;

pub use vad::VadConfig;

pub use vad::VadError;

pub use vad::VoiceActivityDetector;

pub use vad::segment_speech;

Modules

ahc

Agglomerative Hierarchical Clustering (AHC) for speaker diarization.

cluster

Speaker clustering with online incremental centroid updates.

der

Diarization Error Rate (DER) computation.

embedding

Speaker embedding extraction trait.

features

Log-mel filterbank (fbank) feature extraction for speaker embeddings.

offline

Offline (file-based) speaker diarization.

online

Online (streaming) speaker diarization.

overlap

Overlap detection: identify frames where multiple speakers may be active.

pipeline

High-level diarization pipeline.

rttm

RTTM (Rich Transcription Time Marked) parser and writer.

silero_vad

Silero VAD v5 ONNX integration.

types

Core types for speaker diarization.

utils

Math utilities for diarization.

vad

Voice Activity Detection trait and utilities.

wav

WAV file I/O via the hound crate.