ClickClips didn't start with a business plan, a market analysis, or some brilliant idea about disrupting video editing. It started much more simply: I was streaming a lot, mostly Call of Duty: Warzone, and at some point I realized that I had neither the time nor the energy to do anything useful with all the video I was producing.
A normal session for me could easily last three or four hours. When you play the same game regularly, those recordings accumulate surprisingly fast. At first, you think you'll remember where the good moments were. Maybe there was a great fight in the third match, something funny happened an hour later, or you had one of those rare games where everything somehow worked.
A few days later, though, all those matches start blending together. I honestly couldn't remember whether a particular moment happened yesterday, two streams ago, or somewhere in the middle of another four-hour recording. So if I wanted to post something on YouTube Shorts, TikTok or Instagram, I basically had to watch my own stream again just to find the parts worth clipping.
Streaming for four hours is one thing. Watching yourself play the same game for another four hours afterwards is a completely different kind of entertainment.
We tested ClickClips on a 6-hour gaming stream to see how well different highlight detection methods work on long recordings.
We compared AI detection, audio peaks, and speech-based detection, then looked at how many genuinely useful moments each approach found.
One thing that surprised me was that no single detection method works best for every type of moment. Audio is useful for high-energy reactions, speech can find contextual moments, while AI detection can catch visual patterns that the other methods miss.
We wrote up the full test with the results here: AI for highlight detection
