Notepod

AI-powered podcast generator with microservices architecture that creates conversations from various input types.

Architecture

• Spring Boot Server (Java 21) - REST API, authentication, and RabbitMQ publisher
• Python FastAPI Service - AI processing using OpenAI GPT-4 and ElevenLabs for voice synthesis
• React Native Mobile App (Expo) - Cross-platform mobile client with NativeWind styling
• PostgreSQL - Database with persistent storage
• RabbitMQ - Message queue for async processing
• K3s - Kubernetes deployment with Traefik ingress

Features

Podcast Generation Types

• TOPIC - Generate from text topics or prompts
• PDF_DOCUMENT - Analyze and discuss PDF content (with page range selection)
• PHOTO - Create discussions about images using GPT-4 Vision
• WEBSITE_LINK - Scrape and analyze web content

Duration Options

• SHORT (~2 minutes) - Quick summaries
• MEDIUM (~5 minutes) - Balanced discussions
• LONG (~10 minutes) - In-depth analysis

Mobile Features

• Real-time search functionality
• Swipe-to-delete podcast cards (Apple Music style)
• Type-based icons for different content types
• Background audio playback
• Optimistic UI updates with React Query

Quick Start

Prerequisites

• Docker and Docker Compose
• Node.js and npm (for mobile app)
• Java 21 and Maven (for server development)
• Python 3.11 (for AI service development)

1. Environment Setup

Create a .env file in the project root with your API keys:

# AI Services
OPENAI_API_KEY="your-openai-api-key"
ELEVENLABS_API_KEY="your-elevenlabs-api-key"

# AWS S3 Storage
AWS_ACCESS_KEY_ID="your-aws-access-key"
AWS_SECRET_ACCESS_KEY="your-aws-secret-key"
AWS_REGION="ap-south-1"
AWS_S3_BUCKET_NAME="your-s3-bucket-name"

# Clerk Authentication
CLERK_WEBHOOK_SECRET="your-clerk-webhook-secret"
CLERK_JWKS_URI="https://your-clerk-instance.clerk.accounts.dev/.well-known/jwks.json"

# Message Queue (for local development)
RABBITMQ_HOST="localhost"
RABBITMQ_PORT="5672"
RABBITMQ_USERNAME="guest"
RABBITMQ_PASSWORD="guest"

2. Deploy with K3s (Recommended)

Option A: One-command setup

./scripts/k3s-dev.sh

Stop k3 docker

./scripts/k3s-dev.sh down

Option B: Manual steps

# Start K3s cluster
docker-compose -f infra/docker-compose.k3s.yaml up -d

# Wait for K3s to be ready (about 30 seconds)
sleep 30

# Create secrets from .env file
./scripts/create-secrets.sh

# Build and deploy all services
./scripts/build-deploy.sh

3. Access the Services

• API: http://localhost/actuator/health
• API Endpoints: http://localhost/api/podcasts
• RabbitMQ Management: http://localhost:15672 (guest/guest)

4. Mobile App Development

# Install dependencies
cd mobile
npm install

# Start Expo development server
npm start

# Run on specific platforms
npm run ios     # iOS simulator
npm run android # Android emulator
npm run web     # Web browser

Development Commands

Infrastructure Management

# Start K3s cluster
docker-compose -f infra/docker-compose.k3s.yaml up -d

# Stop and cleanup
docker-compose -f infra/docker-compose.k3s.yaml down -v

# View logs
docker exec infra-k3s-server-1 kubectl logs -f deployment/notepod-server -n notepod
docker exec infra-k3s-server-1 kubectl logs -f deployment/notepod-ai -n notepod

# Check pod status
docker exec infra-k3s-server-1 kubectl get pods -n notepod

# Recreate secrets (if .env changes)
./scripts/create-secrets.sh

# Rebuild and redeploy
./scripts/build-deploy.sh

Local Development (Alternative)

Option A: One-command setup

./scripts/local-dev.sh

Option B: Manual steps

# Start PostgreSQL and RabbitMQ only
docker-compose -f infra/docker-compose.yaml up -d

# Run Spring Boot server (Terminal 1)
cd server && ./mvnw spring-boot:run

# Run Python AI service (Terminal 2)
cd podcast-ai && pip install -r requirements.txt && python app.py

# Run mobile app (Terminal 3)
cd mobile && npm install && npm start

Testing

# Spring Boot tests
cd server
./mvnw test

# Python service tests
cd podcast-ai
python tests/test_pdf_podcast.py
python tests/test_image_podcast.py
python tests/test_website_podcast.py

# Mobile app linting
cd mobile
npm run lint

API Examples

Create Topic-Based Podcast

curl -X POST http://localhost/api/podcasts \
  -H "Content-Type: application/json" \
  -d '{
    "type": "TOPIC",
    "sourceContent": "The future of artificial intelligence in healthcare",
    "duration": "MEDIUM",
    "speakers": [
      {"name": "Dr. Smith", "role": "AI Researcher"},
      {"name": "Lisa", "role": "Healthcare Expert"}
    ]
  }'

Create PDF-Based Podcast

curl -X POST http://localhost/api/podcasts \
  -H "Content-Type: application/json" \
  -d '{
    "type": "PDF_DOCUMENT",
    "sourceContent": "https://example.com/research-paper.pdf",
    "duration": "LONG",
    "speakers": [
      {"name": "Prof. Johnson", "role": "Researcher"},
      {"name": "Sarah", "role": "Student"}
    ]
  }'

Troubleshooting

K3s Issues

# Restart K3s cluster
docker-compose -f docker-compose.k3s.yaml restart k3s-server

# Check K3s cluster status
docker exec infra-k3s-server-1 kubectl cluster-info

# View ingress status
docker exec infra-k3s-server-1 kubectl get ingress -n notepod

Secret Issues

# Check if secrets exist
docker exec infra-k3s-server-1 kubectl get secrets -n notepod

# Recreate secrets
docker exec infra-k3s-server-1 kubectl delete secret api-secrets -n notepod
./scripts/create-secrets.sh

Database Connection

# Check PostgreSQL logs
docker exec infra-k3s-server-1 kubectl logs -f deployment/postgres -n notepod

# Check if PVCs are bound
docker exec infra-k3s-server-1 kubectl get pvc -n notepod

logs

# Check notepod-ai logs
kubectl -n notepod logs -f -l app=notepod-ai --all-containers --prefix --tail=1000

SSH port forward for rds and rabbitmq

ssh -L 5432:notepod-postgres.cnmw0e08ykfs.ap-south-1.rds.amazonaws.com:5432 -i ~/.ssh/notepod-key.pem ubuntu@notepod.saheen.dev ssh -L 15672:10.42.0.123:15672 -i ~/.ssh/notepod-key.pem ubuntu@notepod.saheen.dev