wx-voice

Convert audio files between Tencent apps (Weixin / WeChat, QQ) and Silk codec with other general formats such as MP3 and M4A.

中文版

This project is a maintained fork of Ang-YC/wx-voice by AngYC. Many thanks to the original author for the foundational work.

Install

npm install wx-voice --save

No additional compilation step required. Pre-compiled Silk SDK binaries are automatically installed for your platform.

Supported Platforms

| Platform | Architecture | |----------|-------------| | Linux | x64, arm64 | | macOS | x64, arm64 | | Windows | x64, arm64 |

Unsupported platform? Clone the repository and run make -C silk. The library will automatically detect and use the locally compiled binaries.

Prerequisites

FFmpeg is bundled automatically via @ffmpeg-installer/ffmpeg and @ffprobe-installer/ffprobe. No manual FFmpeg installation is required.

CLI Usage

wx-voice <command> <options>

Example:
$ wx-voice decode -i input.silk -o output.mp3 -f mp3
$ wx-voice encode -i input.mp3 -o output.silk -f silk

Commands

Command:
  decode    decode to general audio format
  encode    encode from general audio format
  duration  get duration of audio file

Options:
  -i <input>    input file path
  -o <output>   output file path
  -f <format>   format of the output file
  --bitrate     bitrate of the output file
  --frequency   frequency of the output file
  --channels    channels of the output file

API Usage

const WxVoice = require('wx-voice');
const voice = new WxVoice();

// Error handler
voice.on('error', (err) => console.log(err));

// Decode silk to MP3
voice.decode('input.silk', 'output.mp3', { format: 'mp3' })
    .then((file) => console.log(file));
// Output: "/path/to/output.mp3"

// Encode MP3 to silk
voice.encode('input.mp3', 'output.silk', { format: 'silk' })
    .then((file) => console.log(file));
// Output: "/path/to/output.silk"

// Get duration of a silk file
voice.duration('input.silk')
    .then((seconds) => console.log(seconds));
// Output: 10.24

API

new WxVoice([tempFolder])

| Parameter | Description | | ---------- | ----------- | | tempFolder | Folder for temporary files, defaults to system temp |

Throws if Silk SDK binaries are not found (unsupported platform without local build).

decode(input, output, [options])

Decode a Silk/WebM audio file to a general format (mp3, m4a, wav, pcm, etc.).

Returns Promise<string> — resolves with the output file path on success.

voice.decode('input.silk', 'output.mp3')
voice.decode('input.silk', 'output.pcm', { format: 'pcm', frequency: 16000 })

encode(input, output, [options])

Encode a general audio file to Silk or WebM format.

Returns Promise<string> — resolves with the output file path on success.

voice.encode('input.mp3', 'output.silk', { format: 'silk' })
voice.encode('input.mp3', 'output.silk', { format: 'silk_amr' })  // AMR-compatible header
voice.encode('input.mp3', 'output.webm', { format: 'webm' })      // base64 data URI

duration(filePath)

Get the duration of an audio file in seconds.

• For Silk files: parsed directly from the binary frame structure (no FFmpeg needed)
• For other formats: uses FFprobe

Returns Promise<number> — resolves with 0 if the file cannot be read.

voice.duration('input.silk').then((s) => console.log(s + 's'))

Options

| Parameter | Description | | --------- | ----------- | | format | Output format (silk, silk_amr, webm, mp3, m4a, wav, pcm...), default parsed from output file extension | | bitrate | Bitrate in kbps | | frequency | Sample rate in Hz (e.g. 16000, 24000, 44100) | | channels | Number of channels, default 1 |

File types

Decode: Tested on mp3, m4a, wav, pcm Encode: Tested on silk, silk_amr, webm

Unsupported platform

If your platform is not in the supported list, you can build the Silk SDK locally:

git clone https://github.com/binsee/wx-voice.git
cd wx-voice
npm install
npm run build:silk

The library will automatically detect and use the locally compiled binaries at silk/encoder and silk/decoder.

License

MIT © binsee

Original work MIT © Ang YC