rexsimiloluwah / aiminem

An AI-powered Battle Rapper using GPT-3 and WHisper

AIMINEM

An AI-powered battle rapper using OpenAI Whisper and GPT-3 🚀.

AIMINEM DEMO - Watch Video

Uses

• Python (>3.6)
• Sounddevice: for obtaining the audio recording
• OpenAI Whisper: for transcribing the audio recording
• OpenAI GPT-3: for generating a response to the transcribed audio recording

Requirements

• ffmpeg

How it works?

AIMINEM uses a simple interactive CLI program to record audio from a user, then stores the audio stream in a buffer. This audio stream is transcribed using Whisper, then the transcribed text is used to generate a battle rap response using GPT-3.

Usage

1. Clone the repository

$git clone https://github.com/rexsimiloluwah/aiminem
$cd aiminem

2. Create/Initialize a virtual environment using venv or poetry

• For venv

$python -m venv env
$source env/bin/activate

3. Install the dependencies using pip or poetry

• For venv

$pip install -r requirements.txt

• For poetry

$poetry install

4. Update the .env file with your OPENAI_API_KEY as shown in .env.sample

$cp .env.sample .env
#replace <your_openai_api_key>

5. Run the CLI and follow the instructions to use AIMINEM

• For venv

$python main.py

• For poetry

$poetry run python -m main.py

6. Advanced Usage

To view advanced usage instructions:

$python main.py -h