A comprehensive multilingual climate change assistant powered by Amazon Bedrock Nova and Cohere Command-A models. Supports 20+ languages with intelligent routing, real-time responses, and comprehensive citation system.

• Multilingual Support: 180+ languages with intelligent model routing
• Real-time Chat: Server-sent events for streaming responses
• Smart Caching: Redis-based caching with bypass functionality
• Citation System: Comprehensive source attribution with link validation
• Safety Filters: Advanced off-topic and harmful content filtering
• Export Functionality: Save conversations in multiple formats
• Responsive Design: Mobile-first, accessible UI with dark/light modes

• Frontend: Next.js 14 with TypeScript, Tailwind CSS, static export
• Backend: FastAPI with Python 3.12, serves both API and static files
• AI Models: Amazon Bedrock Nova Pro, Cohere Command-A
• Vector Search: Pinecone for climate knowledge retrieval
• Caching: Redis for performance optimization
• Deployment: Single Azure App Service deployment

# 1. Build the complete application
./scripts/build.sh

# 2. Start production server (serves both API and frontend)
uvicorn src.webui.api.main:app --host 0.0.0.0 --port 8000

# 3. Access the application
# Frontend: http://localhost:8000
# API Docs: http://localhost:8000/docs
# Health: http://localhost:8000/health

# Install Python dependencies
poetry install

# Install frontend dependencies
cd src/webui/app
npm install
cd ../../..

# Start development servers (run in separate terminals)
# Backend:
uvicorn src.webui.api.main:app --host 0.0.0.0 --port 8000 --reload

# Frontend (development mode):
cd src/webui/app && npm run dev

# AWS Bedrock
AWS_ACCESS_KEY_ID=your_access_key
AWS_SECRET_ACCESS_KEY=your_secret_key
AWS_DEFAULT_REGION=us-east-1

# Cohere
COHERE_API_KEY=your_cohere_key

# Pinecone
PINECONE_API_KEY=your_pinecone_key
PINECONE_ENVIRONMENT=your_environment
PINECONE_INDEX_NAME=your_index_name

# Redis (recommended for caching)
REDIS_URL=redis://localhost:6379
# Or use individual settings:
# REDIS_HOST=localhost
# REDIS_PORT=6379
# REDIS_PASSWORD=your_password  # if authentication is enabled
# REDIS_SSL=true  # for secure connections

Redis is essential for caching responses and preventing regeneration of answers. Without Redis persistence, all cached responses are lost when the application restarts.

# Start Redis with proper persistence
./scripts/start-redis.sh

Option 1: System Service (Recommended for VPS/EC2)

sudo systemctl enable redis-server
sudo systemctl start redis-server
# Copy persistence settings from redis.conf to /etc/redis/redis.conf

Option 2: Managed Redis (Recommended for Cloud) Use a managed service that handles persistence automatically:

• AWS ElastiCache for Redis
• Azure Cache for Redis
• Redis Enterprise Cloud
• Upstash Redis

Update environment variables with your managed Redis connection details.

Why this matters: See REDIS_CACHE_FIX.md for details on the cache persistence issue and solution.

Arabic, Bengali, Chinese, Filipino, French, Gujarati, Korean, Persian, Russian, Tamil, Urdu, Vietnamese, Polish, Turkish, Dutch, Czech, Indonesian, Ukrainian, Romanian, Greek, Hindi, Hebrew

English, Spanish, Japanese, German, Swedish, Danish

The application includes comprehensive multilingual testing:

# Run the test suite (when test files are present)
python comprehensive_multilingual_test.py

Latest Test Results: 64.6% overall score

• Language Detection: 87.2% (Excellent)
• Query Rewriting: 96.6% (Outstanding)
• Citations: 60.5% (Good)
• Answer Quality: 57.6% (Acceptable)
• Safety Filtering: 37.1% (Needs Work)

├── src/
│   ├── webui/
│   │   ├── api/           # FastAPI backend
│   │   └── app/           # Next.js frontend
│   ├── models/            # AI pipeline and models
│   ├── data/config/       # Configuration files
│   └── utils/             # Shared utilities
├── scripts/              # Shell scripts (build.sh, start-redis.sh, etc.)
├── docs/                 # Documentation (STARTUP_GUIDE.md, ARCHITECTURE.md, etc.)
└── tests/                # Test suite

• Startup Guide: Detailed setup and deployment instructions
• Architecture: Technical architecture and design decisions
• Azure Deployment: Azure-specific deployment guide
• Redis Cache Fix: Cache persistence configuration
• All Documentation: Full documentation including reports and technical notes

• Input validation and sanitization
• Rate limiting and abuse protection
• Content filtering for harmful/off-topic queries
• URL validation for citations
• CORS protection
• No sensitive data logging

The application uses a single deployment model:

1. Next.js builds to static files (out/ directory)
2. FastAPI serves both API endpoints and static files
3. All traffic goes through the FastAPI server on port 8000

• Language Detection: <100ms response time
• Query Processing: <2s average response time
• Caching: Redis-based with intelligent bypass
• Static Assets: Optimized builds with compression

1. Follow the existing code style and patterns
2. Run tests before submitting changes
3. Update documentation for significant changes
4. Use conventional commit messages

See LICENSE file for details.

Built with ❤️ for global climate education and awareness.