AI Sevima Content Video Generation

Sagala, Putriani Pirma A and Malau, Jeremia Kevin Alexander Jagardo (2026) AI Sevima Content Video Generation. Project Report. [s.n.], [s.l.]. (Unpublished)

Text 5054231027_5025231045-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (1MB) | Request a copy

Abstract

Institusi pendidikan seringkali menghadapi tantangan besar dalam memproduksi video promosi dan pemasaran karena kurangnya tenaga ahli multimedia, keterbatasan anggaran, dan proses produksi yang memakan banyak waktu. Untuk mengatasi masalah tersebut, proyek ini mengembangkan "Sevima AI Content Creator," sebuah platform web berbasis AI yang dirancang untuk membantu institusi pendidikan dalam membuat video promosi secara otomatis dan efisien.
Sistem ini dibangun di atas arsitektur modern tiga lapis yang terdiri dari frontend Next.js, backend Golang, dan layanan AI berbasis Python FastAPI, yang seluruhnya dikemas dalam kontainer Docker dan di-deploy melalui Railway dan Vercel. Alur kerjanya mengintegrasikan Large Language Models (LLM) untuk secara otomatis memproses input dari pengguna (berupa Business Brief dan Creative Brief) menjadi pilar konten, tema, dan storyboard. Untuk rendering akhir, platform ini menggunakan model Google Veo 3.1 Lite melalui Wavespeed AI guna menjalankan proses pembuatan video dari teks (text-to-video generation) berdasarkan deskripsi visual yang telah dibuat.
Berdasarkan pengujian black-box, platform ini mencapai tingkat keberhasilan 100% dari 75 kriteria pengujian fungsional, membuktikan kemampuannya dalam mengelola proses pembuatan video dari awal hingga akhir (end-to-end) dengan lancar. Namun demikian, sistem saat ini masih memiliki keterbatasan, seperti durasi video yang hanya 6-8 detik per scene akibat batasan dari model AI, serta tingginya ketergantungan pada stabilitas API pihak ketiga. Pada akhirnya, Sevima AI Content Creator berhasil mempercepat proses ideasi dan produksi video pemasaran pendidikan sekaligus mengurangi hambatan teknis yang ada.
=======================================================================================================================================
Educational institutions often face significant challenges in producing promotional and marketing videos due to a lack of skilled multimedia personnel, budget constraints, and time-consuming production processes. To address these issues, this project develops the "Sevima AI Content Creator," an AI-powered web platform designed to help educational institutions generate promotional videos automatically and efficiently.
The system is built on a modern three-tier architecture consisting of a Next.js frontend, a Golang backend, and a Python FastAPI AI service, all managed within Docker containers and deployed via Railway and Vercel. The workflow integrates Large Language Models (LLM) to automatically process user inputs (Business and Creative Briefs) into content pillars, themes, and storyboards. For the final rendering, the platform utilizes the Google Veo 3.1 Lite model via Wavespeed AI to execute text-to-video generation based on the generated visual descriptions.
Based on black-box testing, the platform achieved a 100% success rate across 75 functional test criteria, proving its capability to handle the end-to-end video creation process smoothly. However, the system currently faces limitations, such as a restricted video duration of 6-8 seconds per scene due to the AI model's constraints, and a high dependency on the stability of third-party APIs. Ultimately, the Sevima AI Content Creator successfully accelerates the ideation and production of educational marketing videos while reducing technical barriers.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Sevima AI Content Creator, Generative AI / Kecerdasan Buatan Generatif, Text-to-Video Generation, Video Marketing Kampus, Institusi Pendidikan, Large Language Model (LLM), Business Brief & Creative Brief Input, Content Pillar Generation, Storyboard Video, Google Veo 3.1 Lite, Wavespeed AI Integration, Next.js Frontend, Golang & Python FastAPI Backend, Supabase Database & Storage, AI Credit System, Docker Containerization, Black Box Testing, Promosi Kampus,Educational Promotion.
Subjects:	T Technology > T Technology (General) > T385 Visualization--Technique T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing T Technology > T Technology (General) > T58.8 Productivity. Efficiency
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Jeremia Kevin Alexander J. M.
Date Deposited:	11 Jul 2026 13:09
Last Modified:	11 Jul 2026 13:09
URI:	http://repository.its.ac.id/id/eprint/134691

Actions (login required)

View Item