The clipping glossary

Candidate detection

also: moment detection, auto-detection

Candidate detection is the step where software scans a stream and proposes the segments worth clipping, instead of a person scrubbing a timeline to find them. It produces a shortlist, not finished clips — a human still decides what gets posted.

It's the half of clipping that scales badly by hand, because reviewing a list takes minutes and watching a stream takes hours.

Virality score

A virality score is a single number attached to a detected moment, estimating how likely it is to perform as a short clip. It exists to make a shortlist sortable: with fifty candidates from one stream, the ranking is what turns them into a workable queue.

A score is a prediction, not a verdict. Its job is to put the likely winners at the top of your review, not to replace your judgement about what your audience actually wants.

Live clipping

Live clipping is detecting and cutting clips from a broadcast while it is still airing, rather than after the recording is available. The practical difference is when you can post: a moment from hour one is reachable during hour two, instead of after the stream ends and the VOD finishes processing.

VOD scrubbing

VOD scrubbing is the manual method it replaces — dragging through a recorded stream, usually at double speed and watching the audio waveform, hoping to spot where something happened. It is the single largest time cost in manual clipping, and it scales linearly: twice the streams is twice the hours.

Vertical reframe

also: auto-reframe, 9:16 crop

Vertical reframing converts a landscape clip into the 9:16 shape TikTok, Reels and Shorts expect, while keeping the subject in frame. A naive centre crop fails on stream footage, where the webcam is usually in a corner and the action is somewhere else entirely.

Clips channel

A clips channel is an account that posts short highlights of a streamer, often run by someone other than the streamer. It is the most common on-ramp into clipping as paid work: a clipper builds an audience on someone else's content, then either monetises it directly or converts it into a paid arrangement with the streamer.

Clip farm

A clip farm is an operation running many clips channels at once, usually across several streamers, treating clip output as volume rather than craft. The economics depend entirely on cost per clip found, which is why the watching step — not the editing step — decides whether the model works.

Hook

The hook is the opening seconds of a clip, the part that decides whether a scrolling viewer stays. Because short-form feeds autoplay silently, a hook has to work visually before it works audibly — which is why a clip that starts two seconds late often fails despite containing the same moment.

Clipping retainer

A clipping retainer is a recurring arrangement where a streamer or agency pays a fixed monthly fee for a set volume of clips. It is the standard commercial shape of clipping work, and its margin is set by how many hours the clipper spends per delivered clip.

Coverage gap

The coverage gap is the stretch of a stream nobody was watching for clips — the overnight hours, the shift change, the second channel running at the same time as the first. Moments in the gap are not missed because they were judged weak; they are missed because no one was there.