Viralyt · Live public product · Puna Tech
An analytics engine identifying anomalous video performance, computing views-per-hour velocity benchmarks, and extracting high-CTR packaging formulas across YouTube niches.
Views-per-hour velocity & outlier algorithm
Operational transformation
Creators guessed video topics based on old views, producing videos that missed algorithmic momentum and underperformed for months.
Real-time outlier detection surfacing statistical anomalies gaining 5x normal velocity, allowing teams to produce high-demand packaging predictably.
The challenge
Video teams and media agencies spent 15+ hours weekly manually browsing YouTube channels, guessing what thumbnail and topic formats the recommendation algorithm was actively favoring.
The solution
We built a reactive React 19 application powered by YouTube Data API v3 and velocity algorithms that calculate historical baseline averages, isolating videos gaining views 5x-50x faster than channel norms.
Verified impact
Real outcomes delivered for operational teams and end users.
Technology in context
Your operation will be different