A common sign of breathing issues connected to sleep, particularly Obstructive Sleep Apnea (OSA), is snoring. Early diagnosis and treatment depend on the timely and precise identification of snoring episodes. This paper presents SleepEcho, a real-time snoring detection system that makes use of Deep Learning models and embedded audio processing. SleepEcho is a digital microphone that runs on a Raspberry Pi 4. It records audio, extracts 20 Mel-Frequency Cepstral Coefficients (MFCCs), and then uses a Convolutional Neural Network (CNN) and a CNN–Long Short-Term Memory (CNN-LSTM) hybrid model to classify the data. This technology allows for easy-to-use visualization on the web and on mobile devices by sending real-time predictions to a Firebase Realtime Database. From the results obtained, while the CNN model showed more consistent validation performance with 98% accuracy and faster inference time, the CNN-LSTM model obtained 99% training correctness. SleepEcho provides an effective, non-invasive, and affordable solution for real-time snoring detection, supporting at-home health monitoring, and helping clinicians to identify patients who might need additional evaluation for sleep disorders by fusing local processing with cloud-based data synchronization.