Guide

How to turn your own clips into a video, without editing

Most people have the footage already. It is sitting in the camera roll: the walk, the coffee, the drive, the bit where you talked to the camera. What is missing is not material, it is the hour of dragging clips onto a timeline.

The short answer

Describe the video in one or two lines and let an app find the moments. RawCut watches each clip you add once, learning what is shown second by second and what is said. After that you write something like "a day in my life" or "cinematic b-roll by the water", and it cuts a finished vertical video from your own footage.

How it works, step by step

  1. Add clips. Pick a handful from your camera roll. Each one is analysed once and then reused for every video you make.
  2. Say what the video is about. One line for a theme, or one line per scene if you want a specific order.
  3. Get the cut. The AI matches your words to the exact seconds in your clips and assembles them into a 9:16 video.
  4. Adjust if you want. Trim, reorder, add on-screen text, or ask for another cut from the same footage.

Why not stock footage or AI-generated video?

Because it does not look like you. Generated clips and stock libraries produce videos that could belong to anyone, and audiences notice. RawCut only ever cuts from footage you filmed. If a line you wrote has no matching clip, it says so rather than substituting something that nearly fits.

What it is good for

How many clips do you need?

Three is the minimum the app will work with, and it will say so rather than hand you an empty video. Three is enough for a short cut, but the honest answer is that variety matters more than quantity. Ten clips shot in one room give the cutter less to work with than five clips shot in five different places, because every line you write needs somewhere different to land.

A useful rule: if your script has five lines, you want at least five distinguishable settings in your library. Otherwise several lines end up competing for the same footage, and the result repeats itself.

What to write, and what not to

The single biggest factor in how good the cut is turns out to be the sentence you type. Two rules cover most of it.

Describe what can be seen, not what you feel. "Morning light on the kitchen counter" is findable. "The feeling of a fresh start" is not, because nothing in your footage is labelled with a feeling.

Ask for footage you actually filmed. A line like "aerial shot over the coast" fails not because the app is weak, but because there is no drone footage in your camera roll. This is worth saying plainly: a script written for a stock library will disappoint, and a script written from your own life will not.

Good lines look like this:

If you would rather not write at all, the app can draft a script from what is already in your library. That version never asks for footage you do not have, which is exactly why it tends to produce a fuller cut.

What happens to a line nothing matches

It is left out, and you are told. This is a deliberate choice and it is worth understanding, because most tools do the opposite. When a line has no honest match, the alternatives are to insert something that nearly fits or to admit the gap. Something that nearly fits is worse: it looks like a mistake you made, and you have to find it and remove it. A reported gap tells you exactly what to film next.

How long should the video be?

Shorter than you think. A 15 second target with five lines gives each line about three seconds, which is roughly how long a shot holds attention in a feed. Ask for 60 seconds from a small library and the cutter has to stretch: either it repeats footage or it comes back shorter than you asked. Both are honest outcomes, neither is a good video. Start at 15 or 30 seconds and lengthen only when the library can carry it.

One shoot, several videos

This is the part people discover late. The same footage answers different scripts, so one afternoon of filming is not one post. Describe the same trip three ways and you get three genuinely different cuts, because different lines pull different seconds out of the same clips. Filming again is the expensive part; asking a second question of footage you already have costs a minute.

When the cut is not right

Three things fix almost everything, in order of effort:

  1. Swap one clip. Tap the clip in the timeline and the app offers alternatives that fit the same line, so you are choosing between real options rather than searching your whole library.
  2. Rewrite one line. Usually the vague one. Replacing "something nice" with "the harbour at sunset" changes that segment and leaves the rest alone.
  3. Ask for another cut. A re-cut deliberately avoids the clips the previous version used, so you get a different video rather than the same one again.

What it costs, and what leaves your phone

Each clip is read once, and that reading is reused by every video you make afterwards, so the app gets faster and cheaper to use the longer you have it. Your videos are not uploaded. To understand a clip, a small number of downscaled frames and a short audio track are analysed once, and the finished video is assembled on the device.

What it is not

It is not a general-purpose timeline editor, and it does not generate video. If you want frame-by-frame control, keyframed effects, or footage you did not film, this is the wrong tool and a timeline editor is the right one. It also will not rescue footage that does not exist: it can only find what you actually shot.

Related guides

Download RawCut on the App Store

Free to try. iPhone, iOS 15 or later.