AgriGuardian

Inspiration

Millions of farmers around the world lack reliable internet access, yet they make critical decisions daily about weather, pests, and crop care. Inspired by this challenge, we built AgriGuardian, an AI-powered SMS chatbot designed to bridge the digital divide in agriculture. Our goal: deliver smart, accessible guidance to farmers—right from their feature phones.

What We Built

AgriGuardian is a lightweight SMS chatbot that uses Amazon SNS and AWS Lambda to simulate a two-way messaging service. When a farmer sends a question (e.g., “Why are my tomato leaves turning yellow?”), the message is routed to Amazon Bedrock, where a GenAI model analyzes it and returns actionable, localized advice.

We also mocked IoT sensor data (temperature, humidity, soil moisture) to show how AgriGuardian could integrate with low-cost hardware for real-time decision support.

How We Built It

• Backend: AWS Lambda (Node.js) for SMS handling and API integration

• Messaging: AWS SNS to simulate sending and receiving SMS

• AI Engine: Amazon Bedrock (Claude/GPT-J) for language understanding and response generation

• IoT Simulation: Mock sensor data (e.g., temperature, humidity) to test future real-time extensions

What We Learned

• How to connect SMS-based systems with powerful AI models

• Designing for users in low-bandwidth, low-literacy environments

• Mocking IoT data streams to demonstrate potential integrations

• Importance of clarity and trust in AI-generated agricultural advice

Challenges Faced

• Ensuring AI responses were accurate and actionable for real-world farming

• Working around real SMS delivery constraints with mocked SNS triggers

• Balancing language simplicity for rural users with technical accuracy

• Simulating IoT input without real hardware

Built With

• api
• cli
• iot
• openrouter
• python
• requests
• sensors
• simulated
• streamlit