Axion CLI

An AI CLI terminal - similar to Gemini's CLI - compatible with OpenAI-style APIs (supports Open WebUI API as well).

It can natively access the web to perfrom web searches and fetch/scrape URLs.

Installation

Clone the repository to your machine using git clone:

git clone https://github.com/znx-x/axion-cli

Navigate into the cloned repository:

cd axion-cli

Install all the dependencies using pip:

pip install -r requirements.txt

• You will need to use venv and set up a Virtual Environment if you're using macOS or Linux.

Configuration

Rename the config.ini.EXAMPLE to config.ini:

mv config.ini.EXAMPLE config.ini

Edit the configuration file to add your API endpoint and key. It supports any API using OpenAI's standard endpoints, including Ollama and Open WebUI.

nano config.ini

Tool Calling

Axion has Web Search, URL Scraping, and Time Awareness tools built-in. There's no API-keys or any configuration required for these tools to function, but you need to make sure that you set the tool_mode to reflect your model's tool calling framework - either native or manual (for models that do not support OpenAI tool calling).

You can enable or disable access to these tools via the config.ini file.

Running

You can run the application with Python:

python axion.py

To load an existing conversation, you can use the --load flag:

python axion.py --load XYZ

All conversations are saved inside the conversations folder in a .json format.

Compiling

You can compile the code into an executable/binary file with pyinstaller:

pyinstaller --noconfirm --onefile --console --name "Axion" --clean axion.py

If you prefer to use the Axion.spec file to set up your compiler settings, run:

pyinstaller Axion.spec