Salgest

Salgest is an integral AgTech platform designed to monitor grain storage silos in real time, replacing dangerous manual inspections with precision telemetry from ESP32 sensors and visual inspection through onboard cameras. The system reads temperature, humidity, gases, pressure and ultrasonic distance to compute physical stock (level %, tons, dew point, air quality index) and runs a server-side rules engine that generates intelligent alerts for active fermentation, condensation risk, sudden level drops, heat foci, communication loss and more. Each user manages their own silos under a multi-tenant model, with a global gallery, visual history synchronized with sensor timeline, topographic heatmaps and a dashboard of active and historical alerts.

It was built solo as the final project for the Cloud Software Development course at UTN FRLP. The whole stack is deployed on AWS: a Node.js + Express API on EC2 that also serves the React + Vite SPA, PostgreSQL on RDS for relational data (users, silos, sensor_data, alerts, alert configs, gallery captures), and S3 for storing the silo photos sent by the ESP32 (uploaded via multer-s3 with public URLs persisted in the DB). Authentication uses JWT with multi-tenant isolation at every query, the IoT protocol is HTTP multipart/form-data tying each device to a silo via a kit_code, and the frontend renders charts with Recharts and Plotly plus a custom topographic heatmap. The roadmap contemplates a geospatial module with React Leaflet, executive PDF/Excel reports and predictive models (ARIMA / LSTM) to anticipate stock depletion and putrefaction.