An Intelligent Telegram Reminder Assistant powered by Google Gemini and Containerized Architecture.

RemindMe uses a modern Event-Driven Architecture to ensure responsiveness and reliability. It is fully containerized and uses a split-service design.

graph LR
    User -->|Message| Telegram
    Telegram -->|Webhook| Server[Flask Server]
    Server -->|Enqueue| Queue[(Redis Queue)]
    
    Queue -->|Process| Worker[AI Worker]
    Worker -->|Reply| Telegram
    
    Scheduler[Scheduler] -->|Trigger| DB[(Database)]
    DB -->|Reminder| Telegram

• Natural Language Understanding: Just say "Remind me to call Mom in 10 mins" and it understands.
• Asynchronous Processing: Webhooks return instantly; AI processing happens in the background.
• Reliable Scheduling: Dedicated scheduler service ensures reminders are sent on time.
• Cloud Database: Persisted in AWS RDS (PostgreSQL).
• Containerized: Redis and Services are ready for Docker deployment.

• Python 3.10+
• Docker (for Redis)
• Ngrok (for local development)

Create a .env file in the backend/ directory:

TELEGRAM_BOT_TOKEN=your_token
GEMINI_API_KEY=your_key
POSTGRES_DB_URL=postgresql://postgres:password@localhost:5432/remindme
REDIS_URL=redis://localhost:6379
DEV_URL=https://your-ngrok-url.ngrok-free.app

Start the Redis and PostgreSQL containers:

docker compose up -d

You need to run 3 separate terminals to simulate the microservices architecture:

Terminal 1: Web Server (Receives Messages)

source backend/venv/bin/activate
python3 backend/app.py

Terminal 2: Worker (Processes AI)

source backend/venv/bin/activate
python3 backend/worker.py

Terminal 3: Scheduler (Checks Time)

source backend/venv/bin/activate
python3 backend/scheduler.py

This project initially used a serverless architecture with AWS Lambda (for the worker) and Amazon EventBridge (for scheduling).

While we migrated to a containerized worker and scheduler for portability, the original cloud-native code is preserved in the aws/ directory for reference. It demonstrates how to implement:

• AWS Lambda: Serverless execution of Python logic.
• Amazon EventBridge: Cron-style triggers in the cloud.
• AWS RDS: Managed PostgreSQL database for cloud persistence.