EchoScribe - Real-time AI Summarizer with Speech-to-Text Transcription

EchoScribe is a powerful real-time speech-to-text transcription application that leverages Deepgram's WebSocket API for highly accurate and efficient voice recognition.

🚀 Features

• Real-time Speech-to-Text Conversion 📢➡️📝
• AI-powered Speech Recognition by Deepgram 🤖
• Low-latency Streaming via WebSocket 🚀
• Supports Multiple Languages & Models 🌍

📥 Installation

Prerequisites

Ensure you have the following installed before proceeding:

• Node.js (Recommended: v18.x.x or v20.x.x)
• Deepgram API Key 🔑
• Stable Internet Connection 🌐

Setup

Clone the repository and navigate into the project folder:

git clone https://github.com/yourusername/EchoScribe.git
cd EchoScribe

Install dependencies:

npm install

Set up your Deepgram API key:

# Mac/Linux
export DEEPGRAM_API_KEY="your_actual_api_key"

# Windows (CMD)
set DEEPGRAM_API_KEY=your_actual_api_key

Start the server:

node server.js

🎤 Usage

Simply speak into your microphone and see real-time transcriptions displayed in your terminal or web interface.

Modify WebSocket URL parameters to adjust language/model settings as per your requirements.

🛠 Troubleshooting

WebSocket Connection Issues

If you face WebSocket failures, try these steps:

• ✅ Check API Key: Ensure DEEPGRAM_API_KEY is correctly set.
• 🔗 Verify WebSocket URL: Use wss://api.deepgram.com/v1/listen.
• 🌐 Check Internet Connection: Run ping api.deepgram.com.
• 🔄 Downgrade Node.js if needed: Use an LTS version (v20.x.x or v18.x.x).

Debugging Tips

Add logging in server.js for better issue tracking:

console.log("Connecting to Deepgram...");
ws.on("open", () => console.log("✅ WebSocket connected"));
ws.on("close", () => console.log("❌ WebSocket disconnected"));
ws.on("error", (err) => console.error("WebSocket Error:", err));

📜 License

This project is licensed under the MIT License.

👨‍💻 Author(s)

• Pareekshith P
• Arun Kumar K
• Mohammed Naseem
• Pozhilan A

---