EasyCaption

• 

  EasyCaption logo

• 

  gordon ramsay pasta

• 

  spanish playground

Track

I think the project aligns with the "Voice & Conversational AI" track, since its core functionality relies on transcription.

Inspiration

Since I always prefer videos with captions over pure audio, I wanted an easy and fast way to caption videos that often lack them.

What it does

EasyCaption captions videos in English using a speech to text model. This works for videos in any language. For example, a video in Spanish will have English subtitles. The resulting video can be watched directly in the web app or downloaded for later.

How we built it

I used a Vite + React + TypeScript template to get started with the project. I used FastAPI for the backend, and FFmpeg command line executed with Python subprocess to handle the video creation process. Whisper was used for its speech to text transcription and timing.

Challenges we ran into

I had an issue with loading the videos from the backend and displaying them on the frontend. I tackled this issue by assigning a unique key to each video, so that the page would re-render on every new video.

Accomplishments that we're proud of

I am very proud to have gone from nothing to a working product in <7 hours.

What we learned

I learned how to serve files from the frontend to the backend, and vice versa.

What's next for EasyCaption

I want to expand on this project to make it cloud hosted and readily available for anyone to use.

Built With

• fastapi
• ffmpeg
• python
• react
• typescript
• whisper