Guide
How to turn a long video into short videos
When one long recording becomes several short videos, each short video keeps one question, one conclusion, and one example. Cut the greeting, and make the clip understandable without the clips around it.
Updated
What do you decide before you cut?
Write the question the viewer wants answered, then look for the answer in the footage. Without that question, the cuts follow the clock. The clips repeat each other and never land a conclusion.
How many clips can one recording become?
Count passages that can answer one question on their own, not minutes. A 40-minute product explanation can usually become clips like these:
- What problem the feature solves. The conclusion is in the first 15 seconds.
- Which step it removes compared with doing the work by hand.
- One complete example. The clip ends when the example ends.
“Hi, I’m…” is not its own video. If an introduction is necessary, put a short version in the back half of a clip, not at the start.
What structure does each short video use?
- The first sentence states the question or the conclusion.
- Two or three sentences explain why.
- One concrete example names a real object, step, or number.
- One sentence closes the clip and does not open a new question.
TikTok, WeChat Channels, and Bilibili count completion differently. These four steps still hold on each of them: the viewer does not need last week’s episode.
Why can’t the opening be small talk?
Small talk in a long video waits for the viewer to settle. A short video has no such wait. Cut “today we’re going to talk about” and start from a conclusion that has already happened. If the conclusion depends on a noun, say the full noun in the first sentence.
How do you check captions and facts?
- Check names, brands, model numbers, and amounts against the original recording, not against a guess from the caption file.
- Check homophones separately, especially product names.
- For promises such as “always,” “forever,” or “free,” confirm the speaker actually said that.
Speech recognition is good for a first draft. The proofreading pass before publishing stays.
Where does Reeloya sit in this process?
Reeloya marks topic segments, proposes clip candidates, and writes the caption draft. Setting the question, the final choice after the greeting is gone, fact-checking, and uploading stay with the creator. See how Reeloya edits a talking-head video.
When should you not split the video?
If the whole recording is one continuous argument, and every slice would lack a conclusion, keep one long video. Forcing a split only produces lines with no context. The same test is in which part of the job AI video editing does.