Find the moments worth clipping, automatically
This AI video clipper takes one long video and returns short vertical clips, choosing the moments by reading the transcript rather than cutting at fixed intervals. Pick how many clips you want (1–10) and how long each should be (15–90 seconds), and captions can be burned in automatically.
How it works
- 1
Add your video
Pick one from your library or paste an https:// video URL.
- 2
Set count and length
Between 1 and 10 clips, each 15 to 90 seconds. Turn captions on if you want them burned in.
- 3
Let it read and cut
It transcribes the audio, analyses the transcript for the strongest segments, then cuts them.
What you can make
Podcasters
One hour-long episode becomes a week of vertical clips, without you scrubbing the timeline to find the good bits.
Webinar and course hosts
Turn a recorded session into promotional shorts that point back at the full recording.
Streamers
Pull highlights from a long VOD the day after, rather than clipping live and missing moments.
Agency social teams
Give a client ten posting-ready verticals from one interview, captioned and sized for Reels and TikTok.
This AI video clipper takes one long video and returns short vertical clips, choosing the moments by reading the transcript rather than cutting at fixed intervals.
Clip your first video freeSpecs & limits
| Sources | A video from your library, or an https:// URL (YouTube where enabled) |
|---|---|
| Clips per run | 1 to 10 |
| Clip length | 15 to 90 seconds |
| Output format | 9:16 vertical, for Reels, TikTok and Shorts |
| Captions | Optional, burned in, top / centre / bottom |
| Billing | Transcription per audio-minute, one flat analysis fee, captions per clip — the cutting itself is free |
Tips for better results
- Clip a video with clear speech, not one carrying the meaning in visuals. The segments are chosen from the transcript, so a silent montage gives it nothing to rank.
- Ask for fewer clips than you think you need. Five strong clips from an hour beats ten where half are padding — and you can always run it again.
- Use 30–45 seconds for talking-head content. The 90-second maximum suits stories with a setup, but most single points are finished well before then.
- Turn captions on. Feed video is watched muted, and captions are cheaper here than re-editing a clip nobody stopped for.
- Cost tracks the source length, not the clip count — transcription is per audio-minute. Trimming a two-hour recording down to the relevant hour before uploading halves the bill.
Frequently asked questions
How does it choose the clips?
It transcribes the audio and analyses the transcript for the strongest self-contained segments, then cuts those. It is reading what was said, not detecting scene changes.
Can I clip a YouTube video?
A YouTube source is supported where your workspace has it enabled. Uploading the file or pasting a direct https:// video URL always works. Only clip content you own or have the right to use.
What size are the clips?
9:16 vertical, which is what Reels, TikTok and Shorts want. There is no landscape output.
How much does it cost?
Transcription is charged per minute of audio, analysis is one flat fee per job, and captions are charged per clip. Cutting the video costs nothing — that step runs locally. The dialog shows an estimate before you submit.
How long can the source video be?
There is no hard limit, but cost scales with audio length, so the estimate is shown up front — including a worst case for a two-hour video.
Related tools
Resize one video into vertical, square and landscape in a single run — ready for Reels, the feed and YouTube. No re-editing and no re-upload. Try it free.
Turn any podcast clip into a scroll-stopping audiogram with an animated waveform and captions. Square, vertical or landscape. Make your first free.