CLI-Agent

Talk to cli-agent and it will help you to write bash code to solve your problem.

Installation

pip install cli-agent

Configuration

• Set the default model to use.

Based on Litellm , so most of the models are supported (https://github.com/BerriAI/litellm)

cli config default_model <model_name>

• Set the default ask user behavior, when code is about to run it asks user if he wants to run it or not
  • always # ask user always
  • never # ask user never
  • sometimes # ask user when llms suggests to ask

cli config ask_user <always|never|sometimes>

• Set the retry if generation fails

cli config retry_generation <number>

Usage

Commands:

cli : main cli command ci : direct chat with cli

1. Make call through the cli command

cli run 'kill the chrome' --model cerebras/llama3.1-70b

or

ci 'convert this.mp4  to 720p and remove the audio' 

2. If error happens, you can fix it with the following command.

Error-report takes account of the last command it executed, so it will fix the error with the last command

ci ' also flip the video' --model openai/gpt-3.5-turbo --error-report

command line interface

$ cli --help
Usage: cli [OPTIONS] COMMAND [ARGS]...

  CLI AI: Natural Language Command Line Interface

Options:
  --help  Show this message and exit.

Commands:
  config  View or set configuration values
  run     Run the main CLI AI application

$ cli run --help

Usage: cli run [OPTIONS] [USER_INPUT]...

  Run the main CLI AI application

Options:
  --n-hist INTEGER             Number of history items to consider
  --model TEXT                 Model name to use
  -er, --error-report BOOLEAN  Fix the last error with current modifications
  --help                       Show this message and exit.

Recommended models

• cerebras (https://github.com/cerebras) #fastest model inference in the world on llama3.1-70b
• groq (https://github.com/simonw/groq) # 2nd Fastest model inference