This project integrates IoT, Machine Learning, and Large Language Models (LLMs) to develop an advanced Smart Crop Recommendation and Advisory System for precision agriculture. Real-time soil and environmental parameters—including NPK nutrient levels, soil pH, moisture content, temperature, humidity, and rainfall—are continuously monitored using a network of sensors interfaced with Arduino and ESP32 microcontrollers. These values are displayed locally on an LCD module and simultaneously transmitted to a cloud-enabled Flask server.
The backend system utilizes trained machine learning models to analyze incoming sensor data and generate critical predictions such as the most suitable crop, expected market price, soil fertility status, and irrigation requirements. The prediction engine works in combination with domain-specific agronomic logic to estimate potential yield and fertilizer recommendations.
To enhance user interaction, a Large Language Model (LLM)-powered chatbot is integrated into the system. This intelligent assistant allows farmers to ask natural-language questions related to crop suitability, nutrient deficiencies, soil health, irrigation scheduling, sowing time, and weather-based crop performance. The chatbot provides context-aware responses by combining real-time sensor readings, machine learning outputs, and a structured agricultural knowledge base.
By merging real-time IoT monitoring with AI-driven decision support, the system delivers accurate, personalized, and instant agricultural guidance. This empowers farmers to make informed decisions, reduces resource wastage, improves soil management, enhances crop productivity, and brings precision farming within reach for rural communities.
Introduction
Agriculture is the backbone of many developing economies, yet farmers often face challenges such as unpredictable weather, improper crop selection, nutrient imbalances, declining soil fertility, and lack of timely expert guidance. Traditional farming practices often rely on guesswork rather than scientific decision-making, resulting in lower productivity and inefficient use of resources. With the rise of modern technologies, smart farming has emerged as a promising solution to transform conventional agriculture into a more efficient, data-driven, and sustainable process.
This project aims to bridge the technological gap by integrating Internet of Things (IoT) devices, Machine Learning (ML) models, and Large Language Models (LLMs) to create a Smart Crop Recommendation and Advisory System. The system continuously monitors crucial soil and environmental parameters—such as NPK nutrient levels, pH, moisture, temperature, humidity, and rainfall—using sensors connected to Arduino and ESP32 microcontrollers. These real-time measurements offer an accurate representation of the field’s condition, ensuring reliable inputs for further analysis.
The collected data is sent to a Flask-based backend server, where trained ML models process the information to recommend the most suitable crop, analyze soil fertility, estimate irrigation needs, and predict potential market price. To make the system accessible and farmer-friendly, an LLM-powered chatbot is integrated, enabling farmers to interact with the system in simple natural language. The chatbot intelligently answers questions related to crop suitability, fertilizer requirements, sowing time, weather effects, and yield estimation using both the ML predictions and agricultural knowledge base.
By combining IoT sensing, AI-driven analytics, and conversational intelligence, the proposed system provides farmers with real-time, personalized, and actionable insights. This not only supports informed decision-making but also promotes resource optimization, increases crop yield, and contributes to sustainable agricultural practices. Overall, the system represents a major step toward achieving precision agriculture and empowering farmers with accessible smart farming technologies.

OBJECTIVES
1. To measure soil conditions in real time using sensors (NPK, pH, moisture, temperature, humidity, rainfall).
2. To recommend the best crop using a machine learning model based on the sensor data.
3. To estimate crop price and yield to help farmers plan better.
4. To provide farming advice through a chatbot that understands natural-language questions.
5. To display live data on an LCD and send it to the server through ESP32.
6. To help farmers decide fertilizer and irrigation needs automatically from soil readings.
7. To make farming easier and more accurate using IoT and AI technologies.