Traditional methods to quantify egg freshness, including laborious sorting and chemical analysis, have been widely used with significant errors and sample destruction. This study introduces a non-destructive system for detecting egg freshness using near-infrared spectroscopy and machine learning models. The research successfully developed a portable embedded system device using a low-cost multi-spectral sensor and Raspberry Pi 4B microprocessor for non-invasive egg freshness detection. Our non-invasive proposed system for detecting egg quality was created with an affordable price of around 250 US dollar and is a portable device. Some regression methods, such as Multiple Linear Regression and Support Vector Regression, were utilized to predict egg freshness with a coefficient of determination of 0.8. It was also compared with invasive measurements.