Shorts Generator

Inspiration

We are avid consumers of TikTok, who enjoy the TTS videos such as the Biden, Obama, and Trump videos, and the Reddit story videos. Therefore, we decided to make shortsGenerator so that at any time we can have stimulating short-form content at the tip of our fingers.

What it does

Our program will give the user the ability to generate short-form content via Reddit post scraping or user-prompted chatGPT-generated content. The content will be read out loud via a TTS model and then paired with a user-prompted background video scraped from youtube. The final video boasts a caption feature that is synced with the TTS and is saved locally as an MP4 file.

How we built it

We built the program entirely in Python. The video scraper uses PyTube to scrape a video that matches the type of video the user provides. The Reddit scraper uses Praw. To generate the TTS we use Google TTS. To add the captions and TTS to the video we use MoviePy.

Challenges we ran into

We ran into issues formatting the text when placed onto the images. We hoped to fix an issue before submission where the Reddit scraper failed to parse ".." or "..." which resulted in gTTS failing to convert to speech, but we did not have time. We also ran into some issues with Git, but overall things went well. Our team wanted to have the program automatically upload the generated video to YouTube, but YouTube had to review our project and we didn't have the time.

Accomplishments that we're proud of

We are proud of our ability to work separately on different parts and then work together to integrate them at the end. Colin is specifically proud of the documentation and GitHub repo.

What we learned

We learned how to collaborate over Github as we had to deal with a bunch of merge issues. We gained knowledge and experience in working with multimedia files, and utilizing TTS to convert generated text into speech. One of the most valuable lessons we learned during the hackathon was the importance of distributing work and effectively working as a team. Dividing tasks among ourselves allowed us to tackle different aspects of the project simultaneously, maximizing our productivity and efficiency.

What's next for Shorts Generator

Work towards integration with uploading to multiple media platforms - Youtube, TikTok, Instagram
Provide more customizability for video options
Improve the TTS engine to be more realistic

Built With

gtts
moviepy
praw
python
pytube

Submitted to

Beaver Hacks Spring 2023
- Winner 2nd place

Created by

I worked on the video scraping part of the project as well as the documentation. This was my first hackathon so I'm pretty proud of the work I did.

Colin Pannikkat
Hello! I am a fourth-year CS student interested in the intersection between machine learning and environmental sustainability.
I worked on content generation for this project by scraping posts from Reddit and utilizing OpenAI API.

Like my fellow teammates, this is my first hackathon as well and I'm looking forward to participating in more :)

Sarvesh Thiruppathi Ahila
Hello! I'm a Junior at OSU interested in Computer Engineering.
I worked on processing the caption, clip, and tts as one final video.
Like my co-coders this is my first hackathon so I am looking forward to the results.

Ajinkya Gokule
Hey! I am Junior studying Computer Science interested in AI and Music.
I helped with TTS part of the project.

Like my team, I am also a first year CS student and this was my first hackathon. I definitely found some major gaps in my knowledge and look forward to addressing them over the summer and for future events.

David Gesl