AutoDub

• 

  Google Colab Notebook.

Inspiration

I was inspired to create this project out of the frustration that the content creators had to lose their original voice, their vocal identity due to language barrier.

What it does

An AI system that translates/dub any content in any desired language, retaining the voice of the original creator using voice cloning tools and smart TTS and STT models.

How we built it

It was built and debug, all in VS code. I used Vibe-coding and my architectural engineering to create this project. Since it required high performing GPU computation, I pushed the codes into github repo and ran the complete pipeline in Google's Colab free tier. Additionally, the complete project was created free of cost, without spending even a single dollar.

Challenges we ran into

I faced technical walls while creating a smart STT system that can perfectly sync the translation of the original audio into the desired language.

Accomplishments that we're proud of

I am really proud that I created a method for content creators of YouTube through which, they can retain their vocal identify in all the dubbed languages.

What we learned

It was more of execution than that of learning experience. BTW, it's by first time using Notebook based programming.

What's next for AutoDub

Using the winning price from this hackathon, I would like to scale this project further and ship it. Making sure it's widely available for all the content creators on YouTube and other platforms, breaking the language barriers, so that creators can sell more!

No AI was used for writing the submission of the project:)

Built With

• ai
• colab
• data
• google
• models
• python
• sst
• tts