cff-version: 1.2.0 message: "If you use this work, please cite it using the metadata below." type: software title: "Multilingual Audio Intelligence System: Speaker Diarization, Speech Recognition and Translation" authors: - given-names: "Prathamesh Sarjerao" family-names: "Vaidya" orcid: "https://orcid.org/0009-0000-8091-4256" repository-code: "https://github.com/Prathameshv07/Multilingual-Audio-Intelligence-System" license: CC-BY-NC-4.0 date-released: "2025-09-06" abstract: "An end-to-end audio analysis pipeline that combines pyannote.audio speaker diarization, OpenAI Whisper multilingual speech recognition and a three-tier neural machine translation fallback (Opus-MT, Google Translate wrappers, mBART50/M2M-100), with a focus on Indian languages, served through a FastAPI web interface with waveform visualisation and cached demo mode." keywords: - "speaker diarization" - "automatic speech recognition" - "machine translation" - "Indian languages" - "Whisper" - "pyannote.audio"