Skip to content

Latest commit

 

History

History
65 lines (41 loc) · 1.61 KB

README.md

File metadata and controls

65 lines (41 loc) · 1.61 KB
Raw

Introduction

Welcome to please, a powerful and easy-to-use Command Line Interface (CLI) tool to assist working on the command line.

please currently supports OpenAI's gpt-3.5-turbo and gpt-4.

Usage

It has two main modes of operation:

  1. Ask an LLM how to perform a specific task using CLI

     please <some task description>

In this mode, the LLM is asked to return shell command(s) that perform the given task.

Examples:

> please output hello world
echo "hello world"
> please convert test.m4a to mp3
ffmpeg -i test.m4a -acodec libmp3lame -ab 192k test.mp3

  1. Ask an LLM to change some piped input according to a given prompt

     echo "cat dog car house mouse" | please filter animals

In this mode, the piped input is passed to the LLM, together with the please command line as instruction.

This can be used from inside (Neo)Vim: just mark some text with v/V, then pipe through please like "!please complete this".

NOTE all piped in data and every argument to please is sent to the OpenAI API endpoint!

please currently defaults to using gpt-3.5-turbo. To make it use gpt-4, use the --model gpt-4 argument or (better) set PLEASE_MODEL=gpt-4 alongside OPENAI_KEY.

How would you make that permanent?

> please set PLEASE_MODEL=gpt-4 for every new shell
echo 'export PLEASE_MODEL=gpt-4' >> ~/.bashrc && source ~/.bashrc

Installation

cargo install --git https://github.com/kaspar030/please

Make sure to set your OpenAI API key:

export OPENAI_KEY=sk-....

License

please is licensed under the terms of the Apache License (Version 2.0).