Video Sage

Multilingual AI Video Notetaker

01

Problem/Solution

Why did we make our project? What does it solve?

Video Sage - Democratizing Knowledge

• Allows for foreign language speakers to convert from and to not only English, but over a dozen languages from numerous continents
• List includes 26 different languages from numerous continents, truly making the internet a global place for knowledge to spread

02

Technical Walkthrough

Processing Steps

Uses Assembly AI API to transcribe audio, allowing for translation across 2 dozen languages.

Downloading Videos from Youtube using the Youtube API

Transcription

Video Storage

Summarize

Summarizing using Bart, an open-source AI model, to allow for quick note-taking on longer videos

Processing Steps

Utilizes Microsoft Azure‘s Speech Cognitive Services to make dubbed forms of the video’s audio available

Speech to Speech Translation

Language Conversion

Language Conversion using DeepL, which has one of the most comprehensive language models commercially available

Tech Stack

JQuery

Bootstrap

Flask

​

03

Scope

Who does our project help, and to what extent?

Audience

Overcoming language barriers in lectures

Ease of access to the information in different languages

Professionals

Foreign Students

Hearing Impaired

Create quick note documents to share

Quick Attainment of Information

Efficient learning on the go

Teachers

Busy People

04

Future Steps

What will we change in future?

Future Steps

More Languages

More Regional Dialects

Better Notes (highlights, keywords) and Dubbed Videos

More Applications Supported (Vimeo, Privately Hosted Videos, Discord, etc.)

Wider Operating System Support (iOS, Android)

Customizable Summarization Parameters (Bullet Point Length, Detail, etc.)

Live Demo

CREDITS: This presentation template was created by Slidesgo, and includes icons by Flaticon, and infographics & images by Freepik