🚀 Get Response-Lite: Terminal-based AI Assistant

Get Response Lite is a lightweight, modular Node.js CLI tool designed to bring the power of Google's Gemini AI directly to your terminal. Whether you need a quick answer, an interactive chat session, or automated terminal command execution, Get Response Lite provides a seamless and aesthetically pleasing experience.

Created by Swapnoneel Saha, this is the streamlined version of the Get-Response package, optimized for speed and minimal overhead.

✨ Features

• Instant Answers: Get high-quality responses from Gemini 2.5 Flash models instantly.
• Context Awareness: Use files (-f) or entire directories (-d) as context for your questions.
• Interactive Chat: Enter a persistent chat session (-c) that remembers the context of your conversation.
• Terminal Automation: Let AI generate and execute terminal commands (-t) based on your natural language prompts.
• Beautiful UI: Rich text formatting, spinners, and boxed outputs for a premium terminal experience.
• Smart Filtering: Automatically skips large files, binary data, and common build directories (node_modules, .next, etc.) to keep context relevant and within token limits.

🏗 Architecture Overview

The project follows a modular architecture to ensure clean separation of concerns and ease of maintenance.

Directory Structure

get-response-lite/
├── index.js                # Main entry point & CLI router
├── package.json            # Project manifest & dependencies
├── src/
│   ├── ai.js               # Google Generative AI integration & generation cores
│   ├── chat.js             # Interactive chat mode implementation
│   ├── config.js           # Centralized configuration
│   ├── terminal.js         # Command generation & execution logic
│   ├── ui/
│   │   └── help.js         # stylized help message UI
│   └── utils/
│       ├── context.js      # File & Directory reading with smart skipping
│       ├── formatter.js    # Markdown-to-Chalk text styling
│       └── system.js       # OS detection & update checks

Module breakdown

• index.js: The traffic controller. It parses CLI arguments and routes them to the appropriate module (Chat, Terminal, or Simple Ask).
• src/ai.js: Handles the direct communication with the @google/generative-ai SDK. It initializes the model and manages the prompt-response cycle.
• src/utils/context.js: Contains the logic for recursively reading directories and files. It includes a "smart skip" mechanism to avoid reading huge binary files or dependency folders that would clutter the AI's context.
• src/terminal.js: A unique module that translates natural language into shell commands. It includes a safety-first execution loop that asks for user permission before running any generated command.
• src/utils/formatter.js: A custom text parser that converts standard Markdown (like code blocks and bold text) into chalk-styled terminal output, making the AI's response highly readable.

📦 Installation

To use Get Response Lite, ensure you have Node.js installed. Install it globally via npm:

npm i get-response-lite -g

🛠 Usage & Examples

Basic Question

npx ai "How does asymmetric encryption work?"

With File Context

npx ai "Optimize this function" -f ./src/utils/formatter.js

With Directory Context

npx ai "Explain the architecture of this project" -d ./src

Interactive Chat Mode

npx ai -c

Note: You can combine -c with -f or -d to start a chat session with pre-loaded context.

Terminal Automation

npx ai "Find all .js files and count the lines of code" -t

⚙ How It Works (The Logic Flow)

1. Initialization: On startup, the tool checks for updates using latest-version and initializes the Google Generative AI model using a base64-secured API key.
2. Context Gathering: If -f or -d is used, the system reads the targeted files, strips out irrelevant data (using isSkippable), and bundles the content into a context block.
3. Prompting: The user's question is combined with the gathered context and sent to the Gemini model.
4. Formatting: The raw markdown response from the AI is passed through the textFormat utility, which applies colors, italics, and code block styling.
5. Execution (Terminal Mode): If in terminal mode, the AI's response is parsed into individual shell commands. The tool then iterates through these commands, prompting the user for (y/n) permission before each execution.

🤝 Contributing

Contributions are welcome! If you have ideas for new features or bug fixes, feel free to:

1. Open an issue.
2. Submit a pull request with your improvements.

📄 License & Credits

• Author: Swapnoneel Saha
• License: CC-BY-NC-SA-3.0
• Main Package: get-response

Built with ❤️ for terminal lovers.