The Ultimate Desktop Voice Assistant with a Seamless CLI Experience
IRIS-MINI is a production-grade, state-of-the-art Voice AI Assistant. Harnessing the power of the Google Gemini Live API, IRIS provides real-time, conversational interactions combined with powerful system integrations.
Whether you prefer a beautiful graphical interface or a blazing-fast command-line experience, IRIS-MINI adapts to your workflow seamlessly.
| Premium Web Interface | Hacker-Style CLI |
|---|---|
![]() |
![]() |
| Beautiful Three.js Visualizer & React UI | Clean, Silent & Branded Terminal Experience |
- ποΈ Advanced Voice AI: Powered by Google Gemini Live API for real-time, low-latency, and highly intelligent conversational experiences.
- π» Deep OS Integration: Seamlessly open, close, and manage applications on your local Windows machine.
- π Web Search & Knowledge: Instantly search Google and fetch accurate, up-to-date information.
- π CLI Powerhouse: Exposes an easy-to-use Command Line Interface (CLI) for power users to manage the assistant entirely from the terminal.
Built with modern, robust technologies ensuring high performance and a premium feel:
- Frontend: React, Tailwind CSS, Framer Motion, Three.js
- Backend: Node.js, Express, Socket.io
- AI Core: Google Gemini Live API, Glowe Agent
- System: Windows API, Decibri (Microphone integration)
Before you begin, ensure you have obtained the necessary API Keys:
You can use IRIS-MINI as a globally installed CLI (Recommended) or run it locally from the source code for development.
The most robust way to experience IRIS is directly through your terminal.
β οΈ Important Requirement: IRIS-MINI strictly requires Node.js v24+. Please ensure your environment is updated before proceeding.
Install the package globally via npm to make the iris command available everywhere on your system:
npm install -g iris-miniLaunch the assistant for the first time:
irisOn this initial run, an interactive Setup Wizard will launch. It will guide you to:
- Input your Gemini API Key.
- Select your preferred Voice Personality (Choose between the crisp tones of Lyra or Puck).
Note: Subsequent runs will completely bypass this setup, instantly booting the Neural Core and dropping you straight into the Voice AI experience.
If you want to poke around the codebase, tweak the UI, or contribute to the project, follow this path.
git clone https://github.com/201Harsh/IRIS-Mini.git
cd IRIS-MiniEnsure you are on Node.js v24+, then run:
npm installCreate a .env file in the root directory based on the following template. Replace the placeholders with your actual API keys:
# Google Gemini API Key for Voice AI
GEMINI_API_KEY=your_gemini_api_key_here
Start both the backend server and the React frontend concurrently in development mode:
npm run devFor headless environments, homelab servers, or cloud deployments, IRIS-MINI comes fully containerized out-of-the-box.
Note: Running via Docker disables local Windows integrations and native microphone access due to container isolation. It is best suited for network-level access or backend services.
Launch the full stack in detached mode:
docker-compose up -d --buildEncountering issues? Here are the most common hurdles and how to bypass them:
-
"Unsupported Engine" Error during
npm install: You are running an older version of Node.js. IRIS-MINI enforces strict environment rules via.npmrc. You must update to Node.js v24+. -
Voice AI / Microphone Not Working (Windows): Windows may block terminal apps from accessing your microphone by default. Go to Windows Settings > Privacy & security > Microphone and ensure that "Let desktop apps access your microphone" is turned ON.
βββ bin/ # CLI Executable
β βββ iris-mini.ts
βββ data/ # Local data storage
β βββ memory.json
βββ public/ # Static Assets
βββ src/
β βββ client/ # Premium React Frontend
β β βββ assets/
β β βββ utils/
β β β βββ AICore.tsx
β β βββ views/
β β β βββ IrisMini.tsx
β β βββ App.tsx
β β βββ index.css
β β βββ main.tsx
β βββ config/ # Configuration handling
β β βββ dot-env.ts
β βββ server/ # Powerful Node.js Backend
β βββ agent/ # Core Voice Agent
β β βββ iris-voice.ts # Gemini Live API
β βββ constants/
β β βββ StreamConfig.ts
β βββ lib/
β β βββ port-picker.ts
β βββ tools/ # Specialized Agents
β β βββ app-agent.ts
β β βββ browser-agent.ts
β β βββ nexus-agent.ts
β βββ utils/ # Utility Functions
β β βββ memory.ts
β βββ main.ts # Server entrypoint
βββ Dockerfile # Docker Containerization
βββ docker-compose.yml # Multi-container setup
βββ .nvmrc # Node Version Control
βββ .npmrc # NPM Configuration
βββ .env # Environment Configuration
IRIS-MINI is built with a production-ready developer experience in mind. It includes:
- Node Version Management:
.nvmrcand.npmrcto strictly control and match Node/NPM environments. - Containerization: Native
Dockerfileanddocker-compose.ymlsupport for easy isolated deployments. - Code Quality: Enforced linting and formatting rules via Prettier and Commitlint.
- Changelog Automation: Configurations for automated, standardized changelog generation.
We welcome contributions from the community! Whether you want to fix bugs, improve documentation, or add new features, please check our Contributing Guidelines to get started.
Please ensure you also review our Code of Conduct to keep our community approachable and respectable.
Security is a priority. For instructions on how to report vulnerabilities and our security practices, please refer to our Security Policy.
This project is licensed under the terms of the included LICENSE file.
Built with π and powered by AI.

