GPT Tools

A set of prototype Node.js command line tools for OpenAI API.

Prerequisites

Before you run these tools, you need the following:

• Node.js installed on your system.
• An API key from OpenAI.

Installation

Clone the repository and install the dependencies:

git clone https://github.com/ru13r/gpt-tools.git
cd gpt-tools
npm install

ffmpeg is used to work with media files and must be installed on your system.

On macOS it can be installed using Homebrew with the following command:

brew install ffmpeg

Configuration

Create a .env file in the project root and add your OpenAI API key:

OPENAI_API_KEY=your_openai_api_key_here

GPT Tools

PDF to speech converter (pdf2mp3)

Converts PDF documents into audiobooks using OpenAI's text-to-speech service. It processes the text extracted from PDFs and converts it into spoken audio in MP3 format. This tool is ideal for creating audiobooks from written documents, enhancing accessibility, and providing a hands-free way of consuming information.

Features

• Extract text from PDF files.
• Convert text to spoken audio using OpenAI's text-to-speech.
• Save the audio in MP3 format.
• Process large PDFs by splitting text into manageable chunks.

Usage

Run the application using the following command:

node pdf2mp3.js <path_to_pdf_file> -o <output_mp3_file>

For example:

node pdf2mp3.js example.pdf -o example.mp3

How It Works

• Reading PDF: The application reads the PDF file and extracts text using .
• Text Processing: The text is split into chunks to manage size limitations of the TTS API.
• Text-to-Speech Conversion: Each text chunk is converted to speech and stored as MP3 data.
• Combining Audio: All MP3 chunks are combined into a single MP3 file.
• Output: The final MP3 file is saved to the specified output path.

Video and audio files transcriber (transcribe)

Processes audio or video files, transcribing audio using the Whisper-1 model, and saving the transcription outputs to a text file.

Features

• Works with small and large video and audio files of different types.
• Transcribes audio using OpenAI's Whisper-1 model.
• ffmpeg tool is used to process and split files into chunks.

Usage

Run the utility with a command in the following format:

node transcribe.js file.mp4 -o file.txt

How It Works

• The utility first uses ffmpeg to split the input audio or video file into files of approximately 10 MB in size. This ensures each segment fits within API constraints for transcription.
• Each file is then transcribed using the Whisper-1 model.
• The output is saved to a single text file.
• Temporary chunk files are removed.

Text improvement (improve)

Improves text quality using the gpt-4-turbo model.

Features

• Works with larger files. Assumes each paragraph has length not exceeding the API limit.
• Improves the text using OpenAI's gpt-4-turbo model.

Usage

Run the utility with a command in the following format:

node improve.js <text_file> -o <output_file.txt>

How It Works

• The utility first splits the file using newlines.
• Each chunk is then sent to GPT completion API.
• The output is saved to a single text file.