Pengembangan Aplikasi Artificial Intelligence (AI) untuk Konversi Audio dan Video ke Minutes of Meeting (MoM) Menggunakan Large Language Model (LLM)

Ravelia, Rayssa (2025) Pengembangan Aplikasi Artificial Intelligence (AI) untuk Konversi Audio dan Video ke Minutes of Meeting (MoM) Menggunakan Large Language Model (LLM). Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5025211219_Project-Report.pdf - Accepted Version Download (6MB)

Abstract

Pencatatan Minutes of Meeting (MoM) secara manual sering kali tidak efisien dan rentan terhadap hilangnya informasi penting. Untuk mengatasi hal tersebut, dikembangkan MoMify, sebuah aplikasi berbasis AI yang secara otomatis mengonversi rekaman audio atau video menjadi ringkasan MoM yang terstruktur. Sistem ini mengintegrasikan Whisper untuk proses transkripsi dan GPT-4 untuk penyusunan ringkasan, dengan dukungan antarmuka berbasis FastAPI dan Streamlit yang memungkinkan komunikasi frontend–backend yang lancar. MoMify mendukung berbagai format dan durasi file serta menyediakan opsi kustomisasi dokumen seperti jenis huruf, warna, dan bahasa. Hasil evaluasi menunjukkan bahwa sistem memiliki performa optimal, dengan nilai Word Error Rate (WER) yang rendah, Readability Score yang tinggi, serta waktu pemrosesan yang efisien. Dengan kemampuan tersebut, MoMify mampu meningkatkan kualitas dan kecepatan dokumentasi rapat, baik untuk kebutuhan organisasi maupun individu.
======================================================================================================================================
Manual recording of Minutes of Meeting (MoM) is often inefficient and prone to information loss. To address this issue, MoMify was developed—an AI-based application that automatically converts audio or video recordings into structured MoM summaries. The system integrates Whisper for transcription and GPT-4 for summarization, supported by a FastAPI and Streamlit interface to ensure seamless frontend–backend communication. MoMify supports various file formats and durations, and allows users to customize document aspects such as font, color, and language. Evaluation results indicate optimal performance, with a low Word Error Rate (WER), high Readability Score, and fast processing time. These capabilities enable MoMify to enhance the quality and efficiency of meeting documentation for both organizational and individual use.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Artificial Intelligence, Large Language Model, Minutes of Meeting, Natural Language Processing, Speech-to-Text
Subjects:	Q Science Q Science > Q Science (General) Q Science > QA Mathematics Q Science > QA Mathematics > QA336 Artificial Intelligence
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Rayssa Ravelia
Date Deposited:	20 Jun 2025 07:41
Last Modified:	20 Jun 2025 07:41
URI:	http://repository.its.ac.id/id/eprint/119197

Actions (login required)

View Item