How long should a short clip be? What our data from real videos says
By the Clipzi team · Updated
The median clip Clipzi's AI detected in the last 30 days runs 43 seconds, and half of all clips land between 31 and 61 seconds. That comes from 148,852 clips found in 10,589 long videos (6,561 hours of podcasts, interviews, streams and talks) uploaded by creators, not from a lab sample.
Below are the numbers by video length, language and number of speakers, how we measured them, and what they mean when you turn a long video into shorts.
The typical short clip is 30 to 60 seconds
75% of detected clips are 60 seconds or shorter. The middle of the distribution (31 to 61 seconds) is long enough for a complete thought, a setup and a payoff, and short enough for TikTok, Reels and Shorts.
This matches what editors do by hand: very short cuts lose the context that makes a moment land, and clips past two minutes rarely hold attention on vertical feeds. If you want to pick moments yourself, our guide on finding the best moments in a long video explains the signals to look for.
How many clips you get depends on video length
Longer sources give more clips, but not proportionally. Short videos are dense; long conversations have more setup, digressions and repetition per minute.
- Under 10 minutes: median 3 clips per video, median clip 32 seconds (4,021 videos)
- 10 to 30 minutes: median 11 clips per video, median clip 42 seconds (2,428 videos)
- 30 to 60 minutes: median 19 clips per video, median clip 45 seconds (1,706 videos)
- 1 to 2 hours: median 31 clips per video, median clip 45 seconds (1,725 videos)
- Over 2 hours: median 31 clips per video, median clip 52 seconds (709 videos)
Clips get longer as the source gets longer
Median clip length grows from 32 seconds in videos under 10 minutes to 52 seconds in the longest ones. Long-form conversations build ideas slowly, so a complete moment needs more runway. For podcasts, plan for clips closer to a minute than to 30 seconds; see how to turn a podcast into clips.
Who uploads: languages and speakers
61% of the videos are in Spanish and 28% in English. 53% have three or more speakers and 22% have two, which is why speaker-aware framing and subtitles by speaker matter more than single-presenter templates.
The median source video is 19 minutes long, and 1 in 10 runs over 104 minutes.
How we measured this
We took every long video that finished processing in Clipzi in the 30 days before October 2026 and had a transcript duration, and every clip the AI detected in them (clips users created by hand are excluded). Lengths are end minus start time of each detected clip. Numbers are aggregated and anonymous; no video content is published. We refresh this report monthly.
Find the best clips in your next long video
Upload it and get AI-detected moments with titles, ready to reframe and caption.
Try Clipzi FreeNo credit card required