README.md

metadata

title: PensionBot
emoji: 🎤
colorFrom: blue
colorTo: green
sdk: docker
sdk_version: latest
app_file: app.py
pinned: false

PensionBot - Voice Assistant 🎤

A sophisticated AI-powered voice assistant designed for government pension queries and document searches. Built with FastAPI, this backend provides comprehensive API endpoints for voice interaction, document processing, and intelligent responses.

🚀 Features

• Voice Processing: Advanced ASR and TTS capabilities
• Document Search: RAG-based government document knowledge base
• Hybrid AI: Multiple LLM providers for optimal responses
• WebSocket Support: Real-time communication
• Authentication: JWT-based secure access

📡 API Endpoints

• GET / - Service information and available endpoints
• GET /health - Health check with service status
• POST /chat - Text-based conversation interface
• WebSocket /ws - Real-time voice and text communication
• GET /docs - Interactive API documentation

🛠 Technology Stack

• FastAPI: High-performance web framework
• LangChain: AI orchestration and document processing
• LanceDB: Vector database for document search
• Whisper: Speech-to-text processing
• Edge-TTS: Text-to-speech synthesis

🔗 Usage

The API is accessible at the base URL of this space. Use the /docs endpoint to explore the interactive API documentation.

Example Usage:

# Health check
curl https://chabhishek28-pensionbot.hf.space/health

# Chat endpoint
curl -X POST https://chabhishek28-pensionbot.hf.space/chat \
  -H "Content-Type: application/json" \
  -d '{"message": "Tell me about pension policies"}'

📋 Environment Variables

The following environment variables are required:

• GOOGLE_API_KEY: Google Gemini API key
• GROQ_API_KEY: Groq API key for Whisper
• TAVILY_API_KEY: Tavily search API key
• JWT_SECRET_KEY: JWT authentication secret

🔒 Security

This API includes JWT-based authentication for secure access to protected endpoints.

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference