yaoc 0.1.3

Yet Another OpenAI-compatible CLI

Yet Another OpenAI-compatible CLI

A simple, single-file command-line chat client compatible with the OpenAI API.

(or "I just want to quickly test my model hosted with vllm/llama.cpp but don't want to spin up openwebui")

Features

• OpenAI API Compatible: Works with any self-hosted LLM platform that supports OpenAI chat completions API.
• Image Support: Send local or remote images to models that support vision.
• Tools: The script has built-in tools for getting the time, fetching web pages, and performing web searches.
• Syntax Highlighting: Renders Markdown in the terminal for better readability.
• Extremely simple: Single file, no installation needed.

Installation

From AUR

yay -S yaoc-git

From PyPI

pip install yaoc

From Github

1. Clone the repository:

   git clone https://github.com/doryiii/yaoc.git
   cd yaoc
   

2. Install the dependencies:

   pip install -r requirements.txt
   

Or use your operating system's package manager to install python requests, termcolor, rich, and html2text.

Usage

• If installed as system package

  openai-cli --base-url "http://localhost:8090/v1"
  

• If installed with PyPI

  python -m yaoc.main --base-url "http://localhost:8090/v1"
  

• If cloned from Github

  python src/openai_cli/main.py --base-url "http://localhost:8090/v1"
  

Flags

Flag	Description
--base-url	(Required) The base URL of the OpenAI-compatible API.
--model	The model to use. If not given, the first model returned by the API will be used.
--api-key	The API key for the service. Defaults to the OPENAI_API_KEY environment variable.
--system	Optional system prompt to give to the model
--hide-thinking	Hide the thinking process output of the model.
--no-tools	Disable tool calling
--cache_prompt	llama.cpp specific prompt caching

Images

To send an image, use the @image: tag at the end of your prompt:

> Tell me about this image @image:path/to/your/image.png

You can also use a URL:

> What do you see in this image? @image:https://example.com/image.jpg

Piping

You can pipe text into stdin:

echo "Hello!" | python src/openai_cli/main.py --base-url "http://localhost:8090/v1" --hide-thinking

Tools

The script has some built-in tools:

• get_time: Get the current local time.
• web_fetch: Get the content of a webpage.
• web_search: Performs a web search. Requires the LANGSEARCH_API_KEY environment variable to be set.

Dependencies

• requests
• termcolor
• rich
• html2text