Case Study: Multi-Agent AI System To Create Viral Reels For 2M+ Users

Client
Year

2025

Team Size

50-75

User Base

+2M

Technologies
  • Generative AI
  • Cloud
  • Docker
Stack
  • Python
  • Kubernetes
  • React

Creativity & AI come together to create a revolution

Did you know that modern deep learning algorithms can achieve up to 92% accuracy in detecting key moments within videos?

This was a research outcome from an agentic AI project we recently developed for our client onestream.live – a multiplatform recorded streaming web app with millions of monthly active users. They approached us to do develop a multi-agent AI system for detecting and creating viral reel generations from several hours of streamed videos per user.

Our AI-powered tool harnesses this scientific breakthrough, reducing processing times by over 40%—ensuring that every second of your live-stream content is turned into a captivating, shareable highlight.

In today’s fast-paced digital ecosystem, content is king—and speed is crucial. Live-streaming platforms produce hours of raw content daily, yet converting these long recordings into engaging, easily consumable highlights remains a significant challenge. Our solution leverages advanced artificial intelligence to automatically create short, viral reels from long-form live-stream recordings. This dual-purpose innovation was designed for onestream.live’s platform and also launched as a standalone SAAS product.

READ Full Case Study

by Dawood Asjad