🎤 Vosk STT Node.js Bindings

Fast and accurate speech-to-text for Node.js using Vosk

📦 Installation

npm install vosk-stt

🔈 Audio Requirements

Vosk requires mono (1-channel) 16kHz 16-bit PCM WAV files. Convert your audio using FFmpeg:

ffmpeg -i input.wav -ac 1 -ar 16000 -sample_fmt s16 output.wav

| Parameter | Value | Description | |-----------------|-------|------------------------------| | -ac 1 | 1 | Mono audio | | -ar 16000 | 16000 | 16kHz sample rate | | -sample_fmt s16 | s16 | 16-bit PCM format |

🚀 Usage

Basic Recognition

import { runVoskRecognition } from 'vosk-stt';

const result = await runVoskRecognition({
  modelPath: './model/english',
  audioFile: './converted_audio.wav' // Must be mono 16kHz
});

Verify Audio Format

Check your WAV file properties with:

ffprobe -v error -show_streams -select_streams a input.wav

Look for:

channels=1  
sample_rate=16000  
sample_fmt=s16

🧪 Testing

Test files must be in the correct format. Example test setup:

# Convert stereo to mono for testing
ffmpeg -i test/fixtures/original.wav -ac 1 -ar 16000 test/fixtures/test.wav

Then run:

npm test

🛠️ Troubleshooting

Error: "Invalid audio format"

1. Confirm mono channel count:

   ffmpeg -i input.wav -ac 1 output.wav

2. Resample to 16kHz:

   ffmpeg -i input.wav -ar 16000 output.wav

🔄 Batch Conversion Script

Create convert_audio.sh for processing multiple files:

#!/bin/bash
mkdir -p converted
for file in ./input_audio/*.wav; do
  ffmpeg -i "$file" -ac 1 -ar 16000 "./converted/$(basename "$file")"
done

Run with:

chmod +x convert_audio.sh
./convert_audio.sh

This update ensures users understand the strict audio requirements and provides ready-to-use conversion commands. The FFmpeg examples cover:

• Basic single-file conversion
• Audio verification
• Batch processing
• Troubleshooting steps

Would you like me to add any specific advanced audio processing scenarios?