The advent of music platforms with vast libraries has underscored the need for efficient song identification systems. This research proposes a groundbreaking method, the song recognition system via audio fingerprinting, which harnesses unique audio features for accurate song recognition. The system begins by generating a comprehensive local database of audio fingerprints derived from a diverse song library. Employing spectrogram analysis, it extracts distinctive features, hashes, and stores them in the database. During recognition, the system captures input from the microphone and swiftly matches it against stored hashes, providing instantaneous song identification. Integrating chatbot technology and web scraping enriches the user experience by offering detailed song information, including title, artist, genre, and lyrics. By facilitating deeper engagement and exploration of music, this innovative approach bridges modern technology with user simplicity, catering to the diverse needs of music enthusiasts. This research contributes to the evolution of song identification systems by addressing challenges in audio fingerprinting. The proposed system ensures robustness and user-friendliness by leveraging feature extraction-based fingerprint matching and data hashing techniques. Practical spectrogram analysis using Python enhances the system’s efficiency. The system provides comprehensive song details by integrating ChatGPT and web scraping, enriching user interaction and usability. The study advances audio fingerprinting technology and offers promising solutions for more effective and versatile song recognition systems. The proposed system achieves an average accuracy of 96.4% for a 14-s query audio.