Termai

Termai is a lightweight, zero-dependency CLI wrapper for Google's Gemini AI, built for Termux on Android and general Linux environments. It brings the power of Large Language Models (LLMs) directly to your command line, following the Unix philosophy of piping and standard streams.

⚡ Features

• 🚀 Lightweight: Uses standard Python requests. No heavy SDKs or complex dependencies.
• 🟢 Unix Compatible: Supports piping (stdin). Feed logs, code, or text files directly into the AI.
• 🛠 Configurable: Built-in JSON configuration system (ai --config) to edit System Prompts, Temperature, and Models.
• ⚡ Fast: Defaults to gemini-2.5-flash for instant responses.
• 🎨 Clean UI: Minimalist output with syntax-highlighted green text.
• 🧹 Auto-Cleanup: The installer sets everything up and deletes the repository to save space.

📥 Installation

Open Termux and run the following commands:

# 1. Clone the repository
git clone https://github.com/dxsim/litellm_termai

# 2. Enter the directory
cd termai

# 3. Run the installer
chmod +x install.sh
./install.sh

What the installer does:

• Installs Python and required libraries.
• Moves the core logic to a hidden folder (~/.programs/termai).
• Creates a global ai command.
• Self-Destructs: Deletes the downloaded source folder to keep your home directory clean.

🔑 Setup

On the very first run, Termai will ask for your Google Gemini API Key.

• Get a free API key here: Google AI Studio
• Run the command:

  ai "hello"

• Paste your key when prompted. It will be saved locally.

💻 Usage

1. Basic Questions Ask anything directly from the terminal. ai "How do I untar a file in Linux?"

2. Piping (The Power Move) Feed output from other commands into Termai. Debug an error log:

cat error.log | ai "Explain what caused this crash"

Explain a script:

cat install.sh | ai "What does this script do?"

Generate code and save it:

ai "Write a Python hello world script" > hello.py

⚙️ Configuration

Termai comes with a built-in configuration editor. You can change the AI provider, model, and personality. Run:

ai --config

This opens config.json in your preferred editor. The editor is chosen based on the following priority:

1. The $EDITOR environment variable.
2. vim (if installed).
3. nano (as a fallback).

The configuration file looks like this:

{
    "provider": "gemini",
    "proxy": "http://user:pass@127.0.0.1:1080",
    "base_url": "https://litellm_url:4000",
    "gemini_config": {
        "api_key": "YOUR_GEMINI_KEY",
        "model_name": "gemini-2.5-flash",
        "system_instruction": "You are a CLI assistant for Termux...",
        "generation_config": {
            "temperature": 0.7,
            "maxOutputTokens": 1024
        }
    },
    "openai_config": {
        "api_key": "YOUR_OPENAI_KEY",
        "model_name": "gpt-4o",
        "system_instruction": "You are a helpful assistant.",
        "temperature": 0.7,
        "max_tokens": 1024
    }
}

• provider: Set to "gemini" or "openai" to choose your AI provider.
• proxy: (Optional) Set an HTTP or HTTPS proxy for all requests.
• gemini_config: Settings for when provider is "gemini".
  • model_name: Change to gemini-2.5-pro or other available models.
  • system_instruction: Give the AI a persona.
  • temperature: Set to 1.0 for creative answers, 0.1 for precise logic.
• openai_config: Settings for when provider is "openai".
  • model_name: Change to gpt-3.5-turbo, etc.
  • system_instruction: A different persona for ChatGPT.
  • temperature: Controls randomness.
  • max_tokens: The maximum number of tokens to generate.

❓ Help & Troubleshooting

Command List:

ai --help

Re-configure API Keys:

To reset and re-enter your API keys, use the --reinstall flag.

ai --reinstall

Debug Mode:

If the AI isn't responding or you are getting errors, run:

ai --debug "your question"

This will print the raw server response and error codes.

Debug Configuration: If you are having issues with your configuration, you can use the --debug-config flag to print the loaded configuration. API keys will be redacted for security.

ai --debug-config

🗑 Uninstallation

To remove Termai completely:

# Remove the 'ai' command
rm $PREFIX/bin/ai

# Remove the program and configuration files
rm -rf ~/.programs/termai
rm -rf ~/.local/share/termai

3 📄 License

This project is licensed under the MIT License. You are free to use, modify, and distribute this software. See the LICENSE file for more details.

Made with ❤️ for CLI enthusiasts