We Find Video

• 

  WeFindVideo Logo

• 

  PrintScren

Inspiration

Video Generation is the next stage for the AI content Generation. Several times we need to create a content for a Institutional Vodes, Short Video, Reels or a Video Background

What it does

Our backend receive the request to generate the Video Script. Each scene have the following structure:

• SceneNumber
  • Title
  • Description
  • Tags
  • VoiceOver
  • Duration
  • QueryImage
  • Music With the script we do the folling actions: ### Images Query Images with Royalties free to create the scene

Description

Add the Text Description in the Image Botton

VoiceOver

Convert text to speech and add to scence

Music

Search for the more suitable music add to Background

Video Generation

Compose all scenes with the specified duration with Image, text, voice and music apply effect transitions

Upload to Storage

Upload Video Generated to Azure Store Account Container

Render Image

Render Image to Canva Interface

How we built it

React for the Canva Interface Python with Langchain and Moviepy to generate script and Video

Challenges we ran into

Hard to synchronize scenes and overcome API errors that constantly generate bad outputs and need to fix the code

Accomplishments that we're proud of

Create the interesting stack that should evolve in a complete platform

What we learned

Allways need to check the AI output

What's next for We Find Video

Improve User Experience Add Queue Manager to split the video generation and return a better user experience Create a payment for heavy users

Built With

• langchain
• moviepy
• openai
• python
• react