Skip to content

Latest commit

 

History

History
152 lines (114 loc) · 5.87 KB

README.md

File metadata and controls

152 lines (114 loc) · 5.87 KB
Raw

chat-cli

AI Superpowers in your Terminal >_

License

MIT

DISCLAIMER!! This software is currently in an early alpha stage. Use with Caution ⚠️

The application needs your OpenAI API key to generate response. More models coming soon.

We welcome you to take it for a spin, but we do not guarantee the application running fully as per the usage guidelines.

Usage

Basic Usage

There are two base commands available for using the application: chat, and clio.

Coming Soon

  • Create your own specialized agents whom you can invoke in Inline Mode using their individual aliases
  • Open chat-cli in Shell mode to interact with the application in a conversational interface to execute more complex tasks.

Just type chat or clio to see the detailed usage information like shown below:


chat-cli v0.1
AI Superpowers in your Terminal >_

Usage: [ALIAS] [OPTION] [PROMPT]

Aliases:
  chat              The default alias available universally out of the box.
                    Expert in executing tasks in *nix Operating Systems
  clio              The default Agent alias available universally out of the box.
                    Expert in executing tasks in *nix Operating Systems

Options:
  -a, --ask         Ask the user for confirmation before executing any command.
                    Used only after the alias name and before PROMPT along with Safety config set to 'Off'"

* Model        : gpt-3.5-turbo
* Temperature  : 0
* Max. Tokens  : 500
* Safety       : On

Note: The command clio is an acronym for Command Line Input Output, coined to provide a friendly alias to invoke this general-purpose CLI application. You can create your own aliases to invoke the application.

This is the default Inline Mode available out of the box. Check out some of the usage examples below:

clio what is the time now in Paris

clio show me some funny unicode characters

clio show me the top three processes consuming the most CPU

More Usage (Coming Soon)

  • Create your own specialized agents whom you can invoke in Inline Mode using their individual aliases

  • Open chat-cli in Shell mode to interact with the application in a conversational interface to execute more complex tasks.

    Just type chat from your terminal to start chat-cli in an interactive Shell Mode like so:

    chat-cli v0.1 | Using Model: gpt-3.5-turbo | Using Agent: clio | Just Enter your query and Press Enter to get the response. Press Esc for Help, Ctr+C to Exit

Local Development Setup

Initial Setup (Step 1/2)

To begin the setup for the application, first you need to download the project repository from Github:

  • Download the repository:

    git clone git://github.com/adybose/chat-cli.git

  • cd into the repository root:

    cd chat-cli

  • From the repo root, create a python virtual environment and activate it:

    python3 -m venv chatclienv
source chatclienv/bin/activate

  • Install the latest dependencies from the requirements.txt file:

    pip3 install -r requirements.txt

  • Create a config.yaml configuration file from config.yaml.example template file. You can specify which OpenAI model you want to query, and other settings in this file. The safety switch can also be found in this configuration file.

  • Create a .env file from the .env.example file to store environment variables that you want for this application.

Important: OpenAI API Key Configuration

If you have an OpenAI API Key with usage credits, you can configure it for your project on Linux and macOS using any one of these two methods mentioned below:

  • You can either export OPENAI_API_KEY=<yourkey>, or have a .env file in the same directory as main.py with OPENAI_API_KEY="<yourkey>" as a line
  • Add the key to the config.yaml configuration file

Further Setup (Step 2/2)

After fetching the repository and setting up a Python environment with all the dependencies installed, complete the setup of the application using on of the below ways:

1. Setup on Linux and MacOS using the Setup script setup.sh

After downloading and initial setup as mentioned above, simply run the command source setup.sh. This does the following things:

  • Copies the necessary files to ~/chat-cli/
  • Creates two aliases chat and clio pointing to ~/chat-cli/main.py
  • Adds the aliases to the ~/.bash_aliases or ~/.zshrc file

Now (after configuring your OpenAI API key for your application as explained), you can start running the program from the command line like so:

clio what is the time now in Paris

2. Manually Configuring On Linux and MacOS

If you want to manually complete the setup after completing the Initial Setup without using the setup.sh script, follow the steps below:

  • Make the main.py file executable: 
    chmod +x main.py
  • Create an alias to invoke the main.py executable file without using the file name main.py: 
    alias chat=$(pwd)/main.py
alias clio=$(pwd)/main.py #optional
  • Now (after configuring your OpenAI API key for your application as explained), you can call the program from the repo root using the alias followed by your natural language query from the shell like below: 
    clio what is the time now in Paris

CAUTION!! Disabling the Safety Switch

By default clio will prompt the user before executing commands.

To have clio run commands right away when they come back from ChatGPT, change the safety in the config.yaml to false.

If you still want to inspect the command that is executed when safety is off, add the -a argument, Example:

clio -a delete the file test.txt

That's all you need to start using chat-cli on your local Linux or macOS machine.

May the force be with you ⚡

©️ 2023 chat-cli | Made with ❤️ in 🇮🇳