gptcli

gptcli is a command line interface built with Rust that currently offers syntax highlighting, configurable settings, color support, and message history. It offers a user-friendly way to interact with OpenAI's GPT-3 API (and hopefully, GPT-4 aswell!)

Installation

To install gptcli, you'll need to have Rust installed on your system. Once you have Rust set up, you can clone and setup gptcli by running the following command:

git clone https://github.com/kalkafox/gptcli.git
cd gptcli
cargo build --release

This should automatically download and install gptcli on your system.

Usage

To use gptcli after building it, simply run the following command:

./target/release/gptcli

This will start the client in the default configuration.

To use gptcli with a prompt, run the following command:

./target/release/gptcli 'How far away is Andromeda?'

This will initialize gptcli with the specified prompt.

Configuration

To configure gptcli, you can create a config.toml file in the same directory as the gptcli repository. The config.toml file should contain settings in the following format:

[openai]
model = "gpt-3.5-turbo"
temperature = 1.0
top_p = 1.0
n = 1
max_tokens = "inf"
frequency_penalty = 0.0
presence_penalty = 0.0

[app]
prompt = """
Please wrap code in triple backticks, with the language specified. For example, ```python
print('Hello world')
```"""
rainbow_speed = 15.0
notify_save = true
response_prefix = "GPT-3"
rainbow_delay = 100
save_conversation = true

Here's a brief description of each setting:

• OpenAI

  • model: The model to use for the API. Defaults to gpt-3.5-turbo.
  • temperature: The temperature to use for the API. Defaults to 1.0.
  • top_p: The top_p to use for the API. Defaults to 1.0.
  • n: The n to use for the API. Defaults to 1.
  • max_tokens: The max_tokens to use for the API. Defaults to inf.
  • frequency_penalty: The frequency_penalty to use for the API. Defaults to 0.0.
  • presence_penalty: The presence_penalty to use for the API. Defaults to 0.0.

• App

  • prompt: The prompt to use when starting the client. Defaults to `Please wrap code in triple backticks, with the language specified. For example, ```python print('Hello world') ````
  • rainbow_speed: The speed at which the rainbow effect should run. Defaults to 15.0.
  • notify_save: Whether or not to notify the user when a conversation is saved. Defaults to true.
  • response_prefix: The prefix to use for the response. Defaults to GPT-3.
  • rainbow_delay: The delay between each rainbow effect. Defaults to 100.
  • save_conversation: Whether or not to save the conversation. Defaults to true.

Contributing

If you have any issues or feature requests, please open an issue on the GitHub repository. Pull requests are welcome!