@maia-id/maleo v1.1.2

MALEO: Multi Platform Speaker Diarization

A JavaScript library for speaker diarization - the process of partitioning an audio stream into segments according to speaker identity.

Features

• Audio preprocessing with customizable options
• CPU, GPU, WebGPU, and WASM support
• Progress tracking during inference
• Flexible audio input handling
• Silence removal and audio normalization capabilities

Prerequisites

GPU Support

If you plan to use GPU acceleration, ensure you have the required CUDA libraries installed:

libcublasLt.so.12

For CUDA installation instructions, refer to the NVIDIA cuDNN Installation Guide.

Installation

npm install @maia-id/maleo

Usage

Basic Example

import { SpeakerDiarization } from 'speaker-diarization';

// Example usage
const example = async () => {
    const speakerDiarization = new SpeakerDiarization();
    const result = await speakerDiarization.inference({
        audio: './examples/audio.wav',  // File path for Node.js
        language: 'en',
        device: 'cpu', // Device support : 'cpu', 'cuda', 'webgpu', or 'wasm'
        audioOptions: {
            targetSampleRate: 16000,
            normalizeAudio: true,
            removeSilence: true,
            silenceThreshold: -50,
        },
        progress_callback: (progress) => console.log('Progress:', progress)
    });

    console.table(result.segments);
};

example();

Running the Example

node examples/inference.js

Configuration Options

Audio Options

Inference Options

Output Format

The inference method returns a result object containing segments with the following structure:

interface Segment {
    start: number;      // Start time in seconds
    end: number;        // End time in seconds
    speaker: string;    // Speaker identifier
    confidence: number; // Confidence score
}

Citation

If you use this library in your research, please cite:

@inproceedings{irawan2025cross,
  title = {Cross-Platform Speaker Diarization: Evaluating the Scalability of Maleo},
  author = {Eka Tresna Irawan and Ardi Mardiana and Dedy Hariyadi and I Putu Agus Eka Pratama},
  booktitle = {International Conference on Discoveries in Applied Sciences & Advanced Technology 2025},
  year = {2025}
}

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Acknowledgments

• NVIDIA for CUDA support