Interactive-ChatGPT-CLI

An interactive command-line interface for chatting with OpenAI's GPT models, including GPT-4 and GPT-3.5-turbo. Supports various modes with memory for follow-up questions and streaming live responses.

Prerequisites

• Python 3.11 is recommended.
• You need an OpenAI API key. Obtain one from the OpenAI website.

Installation

1. Clone the repository and navigate to the project directory.
2. Install the required dependencies using the following command:

pip install .

Configuration

Add your OpenAI API key in the environment variables, or modify config.py and edit OPENAI_API_KEY to replace it with your API key instead of getting it from the environment. Usage

To run the application with GPT-4, you can use various modes:

Memory mode - (supports follow-up questions but uses more tokens and is slightly slower than non-memory mode.)

With memory and live Streaming response.

gpt --memory-stream

With memory:

gpt --memory

To send a message, press "Esc + Return" while using the chatbot. To reset the memory Inside the session, type !reset-memory and press "Esc + Return"

Non Memory Mode - Doesnt support follow up questions. Best fit for single request and response.

Non memory mode - Streaming response

gpt

Non-memory mode without streaming or live response

gpt --no-stream

To use the GPT-3.5-turbo model, pass the extra argument --gpt3:

gpt --memory-stream --gpt3
gpt --memory --gpt3
gpt --gpt3
gpt --no-stream --gpt3

To exit the session: type !exit and press "Esc + Return"

Known Issues

1. Chat typing animation for non-live response mode is not displaying the time properly.