Meet Lyra AI: How to Setup and Run Your Premium Web Voice Companion 🌌

Welcome to the future of digital companionship. If you are tired of sterile, robotic chat interfaces and want an AI assistant that feels truly alive, warm, and responsive, then it is time to meet Lyra.

Lyra is a next-generation AI voice companion. Unlike generic chatbots, Lyra is prompt-engineered with a distinct personality—she is friendly, emotionally aware, supportive, and communicates naturally in a blend of Hindi and English (Hinglish) or pure English.

Behind the scenes, Lyra combines bleeding-edge technologies like React 19, Tailwind CSS v4.0, Framer Motion, NVIDIA NIM, and Google Gemini 2.5/3.1 to offer ultra-low latency voice-to-voice interaction coupled with a gorgeous, responsive video avatar.

In this guide, we will walk you through what makes Lyra unique, and show you exactly how to set up Lyra on your local machine to begin talking.

---

🔮 What Makes Lyra Different?

✨ Adaptive Video Avatar: Lyra isn’t just a static waveform. She reacts dynamically in real-time, transitioning seamlessly between an idle listening state and an active talking animation using high-definition, dual-layer video sync. 🗣️ Hinglish & English Speech Integration: Lyra speaks and understands English and Hinglish seamlessly. Whether you want a technical partner or an informal friend, she adapts to your tone. 🚀 Hybrid AI Backend: Text Generation: Integrates with NVIDIA NIM (Llama 3.1 8B Instruct) for speed, with auto-fallback to Google Gemini 2.5 Flash for unmatched reliability. Text-to-Speech (TTS): Harnesses the premium Google Cloud TTS voice synthesis (en-IN-Wavenet-D) with a fallback to Gemini 3.1 TTS (gemini-3.1-flash-tts-preview, voice Kore) for warm, natural, and low-latency voice delivery. 🎨 Ambient Mood Effects: Ambient grading, breath glow loops, and glassmorphism styling that react in real-time to Lyra's state (listening, thinking, or speaking).

---

🛠️ Step-by-Step Setup Guide

Getting Lyra up and running on your system is straightforward. Follow these steps to deploy the protocol.

Step 1: Secure Your Lyra Protocol License 🛒 First things first: before running the software, you need to purchase the Lyra protocol license. This gives you the authenticated assets and the core framework code.

👉 Buy the Lyra Protocol here: Falcon AI - Lyra AI Protocol

Step 2: Install Node.js Ensure you have Node.js installed on your workstation. We recommend Node.js v18.x or higher for full compatibility with Vite and ESBuild. Download Node.js: https://nodejs.org/ Verify your installation by running: ``bash node -v npm -v ``

Step 3: Install Project Dependencies Open your command terminal (Command Prompt, PowerShell, or Git Bash), navigate to the folder where you extracted the Lyra repository, and execute the following command: ``bash npm install `` This will install the React, Express, TypeScript, and Tailwind CSS v4 dependencies.

Step 4: Configure the Environment Keys Lyra requires access to an AI API. By default, the application runs on the Google Gemini API. 1. Create a file named .env in the root of the project. 2. Open the file and insert your Google Gemini API key: ``env GEMINI_API_KEY="your_actual_gemini_api_key" ` (Note: You can copy the template from the existing .env.example` file.)

Step 5: Start the Development Server Launch the server with the following command: ``bash npm run dev ` The server will boot up the backend on port 3000` and link up the Vite frontend middleware.

Open your browser and head to: 👉 http://localhost:3000

---

⚙️ Elevating the Experience: Adding Custom API Keys

Once Lyra is running in your browser, look for the Settings (Gear Icon) in the top-right corner. Here, you can insert optional credentials that will be saved 100% locally in your browser cache:

1. NVIDIA NIM API Key: Provides ultra-low latency response times by switching the chat engine to NVIDIA's optimized Llama 3.1 NIM server. 2. Google Cloud TTS API Key: Activates the premium Indian-accented female voice (en-IN-Wavenet-D) for highly natural text-to-speech feedback.

If you don't enter these keys, don't worry! Lyra will automatically fall back to the backend environment's Google Gemini API key for both chat replies and audio synthesis, ensuring you always have a working companion.

---

🎉 Start Talking! Tap the central purple Microphone Button to start speaking. Speak naturally in English or Hinglish. Tap it again to stop listening and let Lyra process, or tap the button while she is speaking to interrupt her.

If you prefer to type, you can toggle the Keyboard Icon next to the mic to enter Typing Mode.

---

Ready to bring Lyra to life? Head over to Falcon AI to buy the protocol and start building your ultimate AI companion today! 💜