chat-cli is a command-line interface (CLI) tool written in Go that allows you to interact with the OpenAI GPT models to generate text from a prompt.

Make sure you have Go installed on your machine. If not, download and install it from here.

You can install chat-cli by cloning the repository and building it:

git clone https://github.com/ruzmuh/chat-cli.git
cd chat-cli
go install .

To use chat-cli, you need to provide a prompt as an argument:

chat-cli "prompt"

You must have an OpenAI API token to use chat-cli. You can obtain it from OpenAI's platform. Once you obtain the token, you must add it to the chat-cli configuration using the --init flag before your first run.

You can also use various flags:

• --codeonly: If this flag is set, only the first code snippet from the response will be printed.
• --init: Initialize a config file for chat-cli. This is especially important for the first-time setup as you need to provide your OpenAI API token.
• --list-models: List all available models and exit.
• --model: Select a specific model to use.
• --token: Provide your OpenAI API token.
• --version: Print the version of chat-cli and exit.

The --codeonly, --model, and --token flags can also be set as environment variables. Prefix them with CHATCLI_:

• CHATCLI_CODEONLY
• CHATCLI_MODEL
• CHATCLI_TOKEN

The priority order for settings is as follows:

1. Command-line arguments
2. Environment variables
3. Configuration file values

This means that if you set a value both as a command-line argument and in the configuration file, the command-line argument will take precedence.

# Use a specific model
chat-cli --model "text-davinci-003" "What is the weather like?"

# Set a model as an environment variable
export CHATCLI_MODEL="text-davinci-003"

# List available models
chat-cli --list-models

# Initialize config
chat-cli --init

# Use a specific token
chat-cli --token "yourtoken" "What is the weather like?"

# Set a token as an environment variable
export CHATCLI_TOKEN="yourtoken"

# Print version
chat-cli --version

chat-cli looks for a configuration file named config.yaml in the following directories, in order:

1. $HOME/.config/chat-cli/
2. /etc/chat-cli/
3. Current directory (.)

You can create or replace the config file in $HOME/.config/chat-cli/ by running chat-cli with the --init command:

chat-cli --init

The configuration file allows you to set the default model and your OpenAI API token.