how-cli

An AI-based CLI assistant to help you with command line & shell.

Demo

how-cli-demo.mp4

Installation

1. Using pip

• Ensure python is installed on your system. (Tested against Python 3.11+)
• Install the package using pip.

pip install -U how-cli

2. Manual Installation

• Clone the repository & cd into it

git clone https://github.com/FireHead90544/how-cli && cd how-cli

• Ensure you're in a virtual environment.
• Install the application.

pip install -e .

Usage

$ how [OPTIONS] COMMAND [ARGS]...

Options:

• -v, --version: Shows the version of the application
• --install-completion: Install completion for the current shell.
• --show-completion: Show completion for the current shell, to copy it or customize the installation.
• --help: Show this message and exit.

Commands:

• setup: Sets up the configuration required to run...
• to: Sends the task to the LLM for analysis.

how setup

Sets up the configuration required to run the application. Set the LLM Provider & the corresponding API Key.

Usage:

$ how setup [OPTIONS]

Options:

• --interactive / --no-interactive: Whether to use interactive mode for setting up configuration? [default: interactive]
• --provider TEXT: The LLM Provider, needs to be passed explicitly if using --no-interactive mode.
• --api-key TEXT: The API Key for the LLM provider, needs to be passed explicitly if using --no-interactive mode.
• --help: Show this message and exit.

how to

Sends the task to the LLM for analysis. Returns the commands to be executed in order to achieve that.

Usage:

$ how to [OPTIONS] TASK

Arguments:

• TASK: The command line task to perform. [required]

Options:

• --help: Show this message and exit.

Providers

how-cli uses ChatModels as they support chat messages as opposed to TextModels and below model providers and their corresponding models are available to use. If you could test the models that are marked as ❌, please create an issue or pull request along with the test results.

Provider	Model	Package	Tested
GoogleGenAI	gemini-1.5-flash	langchain-google-genai	✅
GoogleVertexAI	gemini-1.5-flash	langchain-google-vertexai	❌
GroqMistralAI	mixtral-8x7b-32768	langchain-groq	✅
GroqLLaMa	llama3-70b-8192	langchain-groq	✅
OpenAI	gpt-4o	langchain-openai	❌
Anthropic	claude-3-5-sonnet-20240620	langchain-anthropic	❌

License

how-cil is licensed under the MIT License, it can be found here.

Honourable Mentions

This project is greatly inspired by kynnyhsap's how. Though my implementation is completely different (refer to the below image for architectural details), but at the core both the projects aims to do the same thing. Also, check out LangChain & Typer using which this project was built.