A lightweight AI Command Line Interface powered by OpenRouter API.

This project provides a clean terminal-based AI interaction system with:

• Language selection (English / Indonesian)
• Session memory
• Token limiting
• API key validation
• Configurable model
• Clean terminal UI

• Python 3.9+
• Git
• Internet connection
• OpenRouter API key

This project uses Git for installation across Windows, Linux, macOS, and Android (Termux).

Download from: https://git-scm.com/

git clone https://github.com/jailideaid/WormGPT-V2.0
cd WormGPT-V2.0

pip install -r requirements.txt

python main.py

sudo apt update
sudo apt install git python3 python3-pip -y

git clone https://github.com/jailideaid/WormGPT-V2.0
cd WormGPT-V2.0

pip3 install -r requirements.txt

python main.py

Download from: https://brew.sh/

brew install git python

git clone https://github.com/jailideaid/WormGPT-V2.0
cd WormGPT-V2.0

pip3 install -r requirements.txt

python3 main.py

From F-Droid (recommended).

pkg update
pkg install git python -y
pip install -r requirements.txt

git clone https://github.com/jailideaid/WormGPT-V2.0
cd WormGPT-V2.0
python main.py

When first running the program, it will ask for your OpenRouter API key.

You can create an account on the https://openrouter.ai page and create a key.

The key is stored locally at:

mainAI/.apikey

You can manually edit it anytime.

⚠ Never upload .apikey to GitHub.

Add this to .gitignore:

mainAI/.apikey

After API key verification, you will select:

1. Bahasa Indonesia
2. English

The system enforces response language strictly using system prompts.

All interface text adapts to the selected language.

Edit:

mainAI/config.json

Example:

{
  "base_url": "https://openrouter.ai/api/v1",
  "model": "mistralai/mistral-7b-instruct",
  "site_url": "http://localhost",
  "site_name": "WormGPT CLI"
}

You can change the model anytime.

Each language has its own system prompt file:

mainAI/system-prompts/
    en.txt
    id.txt

When language is selected, the corresponding file is loaded and sent as the system message.

Each request sends:

• System prompt
• Chat history (limited)
• Current user input

To OpenRouter API using:

POST /chat/completions

• Chat history is stored in memory during runtime.
• Limited to prevent token overflow.
• Automatically trims old messages.
• Not persistent after program closes.

The system uses:

• max_tokens for response limit
• History trimming to avoid context overflow

This prevents:

• API errors
• Context length exceeded
• Excess cost

Before entering chat mode:

• The key is validated via OpenRouter auth endpoint.
• If invalid → user is prompted again.
• If valid → stored locally.

If using free-tier models, temporary 429 errors may occur.

wormgpt-cli/
│
├── mainAI/
│   ├── main.py
│   ├── ai_client.py
│   ├── config.json
│   ├── .apikey (automatically creates a file when the key is entered!)
│   └── system-prompts/
│       ├── en.txt
│       └── id.txt
│
└── README.md

• Persistent memory (cross-session)
• Auto-summarization
• Cost tracking
• Multi-model switch
• Plugin architecture

This project is intended for:

• Learning API integration
• Understanding CLI architecture
• Experimenting with language enforcement
• Building AI tools responsibly

Use your API key securely, do not share it publicly.

• don't misuse AI for personal gain to satisfy your crimes.
• You can use wormgpt wisely without having to do or violate anything.
• I am not responsible for that. Thank you for your support by favoriting this repository and following it.

If you use https://openrouter.ai with a free key, if you are hit by a token limit, I suggest changing to another free model or upgrading your key by subscribing.

Built for experimentation and CLI-based AI interaction. Use responsibly.