Generating Minutes of Meeting Using LLM

Abstract

Meetings represent the predominant communication and decision-making medium within organizational frameworks. However, the manual extraction of meeting transcripts from audio recordings is both time-consuming and labor-intensive. This project aims to automate this process by utilizing advanced Large Language Models (LLM) and Speech Recognition techniques through various Python libraries and APIs. The workflow encompasses uploading video or audio content, extracting audio from meeting videos, and transcribing it into textual transcripts, summary and generates minutes via various machine learning models. Sophisticated speaker diarization and identification techniques are applied for segmentation, providing an individual understanding of each speaker's contributions during the meeting, thereby enhancing contextual richness. Furthermore, the project generates concise summaries of the meeting, distilling key insights, action items, and decision points. By automating transcription, speaker segmentation, and summarization, this project streamlines the extraction of valuable insights from meeting audio and video recordings. The resulting transcripts, summaries, and keynotes facilitate better information sharing, knowledge retention, and informed decision-making within organizations, bridging the gap between unstructured audio data and actionable insights.

.