TranscribeCLI - The Seamless Audio-to-Text Whisperer

Table of Contents

• Introduction
• Features
• Requirements
• Installation
• Usage
• Options
• Configuration
• Examples
• Troubleshooting
• License

Introduction

TranscribeCLI is a simple and powerful utility for converting audio files into text, allowing you to seamlessly transcribe .m4a audio files into .wav and process them using OpenAI's Whisper ASR model. This tool is designed to be cross-platform and user-friendly, providing flexible options to suit various use cases.

My Use Case

I prefer to listen attentively in class and take notes myself, rather than just record everything verbatim. But sometimes, things slip through, or I just want to revisit something later to reinforce what I learned. That’s where TranscribeCLI comes in handy.

I record notes or lectures on my phone, Airdrop them to my laptop, and then run this script. The audio is quickly converted to text, so if I’ve forgotten anything or need to go over certain details again, the transcript is ready for me to work with. This way, I can focus on understanding in class and use the transcripts as a backup to ensure I don’t miss any important points.

Table of Contents

• Introduction
• Features
• Requirements
• Installation
• Usage
• Options
• Configuration
• Examples
• Troubleshooting
• License

Introduction

TranscribeCLI is a simple and powerful utility for converting audio files into text, allowing you to seamlessly transcribe .m4a audio files into .wav and process them using OpenAI's Whisper ASR model. This tool is designed to be cross-platform and user-friendly, providing flexible options to suit various use cases.

Features

• Automatic Audio Conversion: Converts .m4a files to .wav format with ease.
• OpenAI Whisper Integration: Transcribes audio files to text using Whisper, supporting quick speech-to-text conversion.
• Cross-Platform Support: Runs on both Linux and macOS, with smart detection for OS-specific configurations.
• Easy Setup: Automatically downloads required binaries (ffmpeg, ffprobe, and Whisper) as needed.
• Symbolic Linking: Provides an option to make the script accessible system-wide without hardcoding paths.
• Customizable Downloads Folder: Choose your preferred directory for storing audio files.

Requirements

Before using TranscribeCLI, ensure that the following build dependencies are installed:

• git: For cloning the Whisper repository.
• make: For building the Whisper executable.
• gcc and g++: Required C and C++ compilers for building.

Installing Dependencies

Ubuntu/Debian

sudo apt update
sudo apt install -y build-essential git

macOS

1. Install Xcode Command Line Tools:

xcode-select --install

2. Install Homebrew (if not installed):

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

3. Install Make and GCC:

brew install make gcc

Installation

Step 1: Clone the Repository

git clone https://github.com/yourusername/TranscribeCLI.git
cd TranscribeCLI

Step 2: Make the Script Globally Accessible (Optional)

To make the transcribe.sh script available globally, create a symbolic link in /usr/local/bin:

./transcribe.sh -p

Usage

To use TranscribeCLI, simply run the transcribe.sh script with the desired option:

./transcribe.sh [option]

Options

• -a: Convert all .m4a files in the Downloads folder to .wav.
• -c: Transcribe all .wav files using the Whisper model and output transcriptions as .txt.
• -w: Convert .m4a files to .wav and then transcribe them.
• -p: Create a symbolic link to make the script accessible from anywhere in your system.
• -h: Display the help message and exit.

Configuration

Custom Downloads Folder

The default location for audio files is ~/Downloads. You can customize this by setting the DOWNLOADS_FOLDER environment variable:

DOWNLOADS_FOLDER="/path/to/your/folder" ./transcribe.sh -w

Setting Up a Symbolic Link for Global Access

After running ./transcribe.sh -p, the script will be accessible from any directory:

transcribe -a

If you encounter a "command not found" error, ensure /usr/local/bin is included in your PATH by adding the following line to your shell configuration file (e.g., .bashrc or .zshrc):

export PATH="/usr/local/bin:$PATH"

Restart your terminal or source the configuration file:

source ~/.bashrc   # For bash users
source ~/.zshrc    # For zsh users

Examples

Convert .m4a Files to .wav

transcribe -a

Converts all .m4a files in the ~/Downloads folder to .wav format.

Transcribe .wav Files

transcribe -c

Processes all .wav files in the ~/Downloads folder using the Whisper model and generates .txt transcriptions.

Convert and Transcribe in One Step

transcribe -w

Converts .m4a files to .wav and then transcribes them into text.

Troubleshooting

Build Issues for Whisper Executable

If the script prompts you to download and build whisper.cpp, but the build fails, try the following:

1. Check Dependencies: Ensure git, make, gcc, and g++ are installed.
2. Permission Issues: Ensure write permissions for the directories being used. Use sudo if necessary.
3. Manual Build: If the automated build fails, manually clone and build the whisper.cpp repository:

   git clone https://github.com/ggerganov/whisper.cpp.git
   cd whisper.cpp
   make

   Move the main executable to the project directory:

   mv main ../TranscribeCLI/

ffmpeg Issues

If ffmpeg and ffprobe are not found:

• Confirm with y when prompted by the script to download these binaries.
• If the download fails, manually download ffmpeg from here.

License

This project is licensed under the MIT License. See the LICENSE file for details.