NIA Backend

The NIA Backend powers the NIA language learning assistant with real-time audio processing and AI integration. Built with Go and optimized for low-latency communication, it handles all AI interactions and audio streaming between the mobile app and OpenAI's services.

Features

• WebSocket Audio Streaming
• OpenAI API Integration
• Authentication Middleware
• Low-latency Processing
• Cloud Deployment
• Self-hosted AI Models

Project Structure

.
├── config/           # Configuration management
├── pkg/
│   ├── api/         # HTTP/WebSocket handlers
│   ├── openai/      # OpenAI API integration
│   └── util/        # Helper functions
└── migrations/      # Database migrations

Getting Started

Prerequisites

• Go 1.19+
• OpenAI API key
• Google Cloud SDK

Installation

# Clone the repository
git clone https://github.com/bielcarpi/nia_backend.git

# Install dependencies
go mod download

# Run locally
go run main.go

Deployment

# Deploy to Google Cloud App Engine
gcloud app deploy

Architecture

The backend implements a clean architecture pattern with:

• WebSocket handlers for real-time communication
• OpenAI service integration (GPT-3.5, Whisper, TTS)
• Authentication middleware
• Helper utilities

Audio Processing

• Real-time streaming via WebSocket
• AAC codec support
• Optimized for low latency
• Future PCM16 support planned