SpeechLingo AI

We are Bridging the language gap in education by translating videos using AI/ML in Local Communities Native Language

Comment

Project Submitted in Advanced Track

Inspiration The inspiration for SpeechLingo came from the realization that a vast amount of educational content is only available in English, which can be a barrier for people who do not speak English in Rural Areas as a first language. We wanted to make this content more accessible by translating it into native languages so that people could learn and grow regardless of their language skills.

What it does SpeechLingo is a platform that allows users to translate English videos into their native language. The platform uses a combination of automatic speech recognition (ASR), machine translation (MT), and text-to-speech AI models to produce high-quality translations. Users can upload videos to the platform, select the desired translation language, and receive a translated video with translated audio in the selected language.

How we built it Front End: NextJS, Backend: IBM-Watson, NodeJS, DeepGram AI, Python

WorkFlow:

  1. We detach audio from video
  2. We use DeepFram's Speech-To-Text AI model to get Speech Data
  3. We use IBM-Watson Language Translator AI model to translate it to native language
  4. We use the IBM-Watson Text-To-Speech AI model to synthesize translated speech into audio
  5. We sync back the Video and send the link of video hosted locally to user

Challenges we ran into: We Faced Issues as More of the native languages weren't supported and Even If they were supported, computer was not able to understand its text

Accomplishments that we're proud of We are proud of the fact that we have been able to create a platform that makes educational content more accessible to people who do not speak English as a first language. We are also proud of the fact that we have been able to build a team of talented engineers who are committed to making the world a more inclusive place.

What we learned On Tech Side: We learned how to work and sync multiple AI models to one on Intellectual Side: We have also learned that there is a great need for platforms that make educational content more accessible to people who do not speak English as a first language.

What's next for SpeechLingo AI We are working on improving the accuracy of our translations, and we are also working on expanding the number of languages that we support.

Share this project:

Updates