ColdLine

ColdLine is a speech-to-text and translation web application that leverages Azure AI services to provide real-time transcription, translation, and AI-powered insights. The platform is built using Flask, JavaScript, and Tailwind CSS, integrating APIs for speech recognition, language translation, and AI-generated responses.

Features

• Speech-to-Text: Converts spoken words into text using Azure Speech Services.
• Language Translation: Supports multiple languages via Azure Translator.
• AI Guidance: Generates AI-powered insights based on user inputs.
• File Uploads: Allows users to upload audio files for transcription.
• Real-time Updates: Displays live transcripts and AI responses.

Tech Stack

• Backend: Flask (Python)
• Frontend: HTML, Tailwind CSS, JavaScript
• APIs Used:
  • Azure Speech Services
  • Azure Translator
  • Azure OpenAI
  • Azure Storage
• Deployment: Uvicorn (ASGI server)

Setup Instructions

1. Clone the Repository

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

2. Install Dependencies

pip install -r requirements.txt
pip install python-dotenv
pip install --upgrade openai httpx
pip install "uvicorn[standard]" websockets wsproto
pip install azure-ai-translation-text
pip install --upgrade azure-ai-translation-text
pip install azure-core azure-ai-translation-text

3. Configure Environment Variables

Create a .env file in the project root and add the following keys:

AZURE_SPEECH_KEY=your_speech_key
AZURE_SPEECH_REGION=your_speech_region
AZURE_TRANSLATOR_KEY=your_translator_key
AZURE_TRANSLATOR_REGION=your_translator_region
AZURE_OPENAI_API_KEY=your_openai_key
AZURE_OPENAI_ENDPOINT=your_openai_endpoint
AZURE_OPENAI_DEPLOYMENT=your_openai_deployment
AZURE_STORAGE_CONNECTION_STRING=your_storage_connection_string
AZURE_STORAGE_CONTAINER=your_storage_container
INITIAL_DATA_BLOB=initial_data.json

4. Run the Application

uvicorn app:app --reload

The application will start on http://127.0.0.1:8000/.

File Structure

ColdLine/
│── static/
│   ├── css/
│   │   ├── styles.css
│   ├── images/
│   │   ├── img1.jpg
│   ├── js/
│   │   ├── app.js
│   ├── index.html
│── app.py
│── requirements.txt
│── .env (not included in repo, must be created)
│── initial_data.json

Project Demo

1. Home Menu

This is the entry point for users to interact with the ColdLine Assistant and explore its features.

1. ColdLine - Hindi Language Support

This screen demonstrates AI-driven responses in Hindi.

• User Query: "आपके पास कौन सी सिक्योरिटी मेजर है?"
• AI Response: Lists enterprise security measures.
• Features: Language selection, voice input, file upload.

---

2. ColdLine - English Language Support

This screen displays AI responses in English.

• User Query: "Why should we use your CRM over others?"
• AI Response: Highlights CRM features, AI analytics, and 24/7 support.
• Features: Real-time interaction, intuitive UI.

---

3. ColdLine - German Language Support

This screen showcases multilingual AI guidance in German.

• User Query: "Warum CRM gegenüber anderen verwenden?"
• AI Response: Mentions CRM integration with Salesforce & Slack, boosting productivity.
• Features: Multilingual support, dynamic AI guidance.