We present a comprehensive pipeline for processing Kannada audio, leveraging state-of-the-art pre-trained models from Hugging Face. The pipeline is designed to handle multiple stages of audio processing, including denoising, speech segmentation, automatic speech recognition (ASR), punctuation, transliteration, translation, and grammar checking. The novelty of our approach lies in the seamless integration of these models into a user-friendly interface, allowing users to upload or record audio and receive outputs at each stage with the option to edit and classify the results.
