Terminal-chat

A simple terminal-based chat application powered by the nginr-preprocessor. This project allows you to interact with AI models directly from your terminal.

Features

• Chat with AI models in your terminal
• Command-based interface
• Easy to extend and customize

Quick Start

Prerequisites

• Python 3.10 or later
• nginr installed in your environment (see requirements.txt)

Installation

1. Clone this repository:

   git clone https://github.com/nginrsw/terminal-chat.git
   cd terminal-chat

2. (Optional) Create and activate a virtual environment:

   python3 -m venv ENV
   source ENV/bin/activate

   you can use any virtual environment tools you have / favorite

   Ignore this command to install all packages from requirements.txt globally (without a virtual environment).

3. Install dependencies:

   pip install -r requirements.txt

Usage

To start the chat application, run:

nginr src/main.xr

To run the tests:

nginr tests/test_main.xr

Configuration Example

This project uses a YAML config file (nginr_config.yaml) to extend and set the model or other parameters. By default, it uses the Deepseek model via OpenRouter:

model: deepseek/deepseek-r1-0528:free
provider: openrouter
temperature: 1.0
top_p: 1.0
max_tokens: 1024

You can change these values in nginr_config.yaml to use a different model or adjust parameters as needed.

Project Structure

• src/ : Source code for the chat application
• tests/ : Test scripts
• docs/ : Documentation (API, architecture, developer notes)
• requirements.txt : Python dependencies

Contributing

Contributions are welcome!

License

This project is licensed under the MIT License.