qprom a.k.a Quick Prompt - ChatGPT CLI

A Python-based CLI tool to quickly interact with OpenAI's GPT models instead of relying on the web interface.

Table of Contents

1. Description
2. Installation
3. Setup
4. Usage
5. Todos
6. License

Description

qprom is a small project that lets you interact with OpenAI's GPT-4 and 3.5 chat API, quickly without having to use the web-ui. This enables quicker response times and better data privacy

Installation

pip install qprom

Setup

Make sure you have your OpenAI API key.

When running qprom the script tries to fetch the OpenAI API key from a credentials file located in the .qprom folder within the user's home directory. If the API key is not found in the credentials file, the user is prompted to provide it, and the provided key is then stored in the aforementioned credentials file for future use.

Usage

Usage

qprom -p <prompt> -m <model> -t <temperature> -v -c

• <prompt>: Replace with your prompt
• <model>: Replace with either gpt-3.5-turbo or gpt-4
• <temperature>: Replace with a float value between 0 and 2
• -v: Add this flag to enable verbose mode
• -c: Add this flag to enable conversation mode

For example:

qprom -p "Translate the following English text to French: '{text}'" -m gpt-4 -t 0.7 -v

This will run the script with the provided prompt, using the gpt-4 model, a temperature of 0.7, and verbose mode enabled.

Multi line prompting

To facilitate multi-line input for the prompt, invoke qprom without utilizing the -p parameter. This will prompt you for your input at runtime, where you can provide multiple lines as needed. To signal the end of your input, simply enter the string 'END'.

qprom

This will run qprom with default values model: gpt-3.5-turbo, a temperature of 0.7 and ask for the prompt during runtime.

Set default model

qprom -M <model-name>

Set token limit for prompt/conversation

qprom -tk <token-limit>

Set default token limit

qprom -TK <token-limit>

Piping console input into qprom

Just pipe the prompt into qprom.

cat prompt.txt | qprom

Todos

• Cleanup project / refactoring
• Add option to set default temperature
• Add option to re-set the API token
• Testing
• Add option to disable streaming and only print the full response

Bug reports:

License

MIT Link

Support me :heart: :star: :money_with_wings:

If this project provided value, and you want to give something back, you can give the repo a star or support by buying me a coffee.