Agri Assist: AI Based Village-Level Crop Risk and Profit Prediction using Voice Assistant.

This paper presented AgriAssist AI, a full-stack web application that integrates MERN-stack web technology with a Python-based machine learning microservice to deliver crop risk classification and profit estimation for village-level farmers. The system is distinguished by its multilingual voice assistant supporting English, Kannada, Hindi, and Marathi; its real-time notification and farmer- admin communication infrastructure built on Socket.io; and its role- based dual-interface design serving both farmers and system administrators. The Random Forest crop risk classifier achieved eighty-seven percent accuracy on a synthetic held-out test set, while the Linear Regression profit estimator achieved an R² of 0.83. End-to-end prediction latency averages 1.8 seconds, making the system practically responsive for real-world use. The work demonstrates that contemporary web frameworks and machine learning libraries can be composed into a coherent, accessible, and feature-rich agricultural decision support platform without requiring specialized infrastructure. By prioritizing accessibility for low-literacy rural users through voice input, local language support, and simple step-by-step interfaces, AgriAssist AI moves toward the goal of equitable access to data-driven agricultural advisory for smallholder farming communities. Real-world field validation and continuous model improvement through operational data remain the immediate priorities for future development.  

[1]      Kulkarni, N.H., Srinivasan, G.N., Sagar, B.M., and Cauvery, N.K., "Improving Crop Productivity Through A Crop Recommendation System Using Ensemble Machine Learning Technique," Proceedings of the 3rd International Conference on I-SMAC, IEEE, pp. 1004–1008, 2020. DOI: 10.1109/I-SMAC49090.2020.9243542 [2]   Van Klompenburg, T., Kassahun, A., and Catal, C., "Crop Yield Prediction Using Machine Learning: A Systematic Literature Review," Computers and Electronics in Agriculture, Vol. 177, 105709, 2020. DOI: 10.1016/j.compag.2020.105709 [3]   Huang, L., Liu, Y., Huang, W., Dong, Y., Ma, H., Wu, K., and Guo, A., "Combining Random Forest and XGBoost Methods in Detecting Early and Mid-Term Winter Wheat Stripe Rust Using Canopy Level Hyperspectral Measurements," Agriculture, Vol. 12, No. 1, 74, 2022. DOI: 10.3390/agriculture12010074 [4]      Ngo, T.T., Assaf, A.G., and Tsionas, M., "Machine Learning Based Agricultural Price Forecasting for Major Food Crops in India Using Environmental and Economic Factors," Biological and Life Sciences Forum, Vol. 54, No. 1, 7, 2025. DOI: 10.3390/blsf2025054007 [5]       Hase, S., and Nimbhore, S., "Speech Recognition Techniques in Agriculture Sector: A Comparative Study," Proceedings of ICIDSSD 2022, EAI, 2023. DOI: 10.4108/eai.24-3-2022.2318772 [6]   Subramanian, V., Pro MERN Stack: Full Stack Web App Development with Mongo, Express, React, and Node, 2nd ed.,Apress / Springer, 2019. ISBN: 978-1-4842-4391-6 [7]   Sindhur, N.M., Pavithra, C., and Muchikel, N., "A Hybrid Machine Learning Framework for Optimizing Crop Selection via Agronomic and Economic Forecasting: A Voice-Enabled Decision Support System for Farmers in Karnataka," arXiv preprint, arXiv:2507.08832, 2025.