Captions that land on the word you just said.

The editor listens to the speech in your clip and writes it down word by word, with a start and an end for every word. Captions are built from those words, so a caption does not drift away from the voice and you are not nudging a block of text against the sound by eye.

That reading step costs nothing. It happens once for each clip, not once for each change you make afterwards.

The words are yours to correct. Open a line, change a wrong word, and the timing stays where the voice put it. If a caption breaks in the wrong place, move the break. If a word sits on the wrong line, send it to the line it belongs on.

Start from a ready-made look, or build your own. You choose the typeface from a short list of heavy display faces and plain readable ones, the weight, and whether the text reads in sentence case, capitals or lower case.

Outline and shadow each come in four steps, from none to large, so you can keep white words readable over a bright wall without guessing at a number. The size has a floor and a ceiling that keep a caption legible and keep it inside the frame.

Then choose how the words arrive. They can cut in with the line, fade in with it, pop in one at a time as each is spoken, show the whole line with a moving highlight behind the word being said, or fill a bar across each word for exactly as long as it is spoken. The highlight can be a box, a pill or an underline.

You set how many words show at once, from one at a time to a short line, and drag the captions anywhere between the top and the bottom of the frame. They stop short of the very edges, so nothing ends up off the picture.

When it reads right, export. The captions are burned into the picture, so they show on every platform and every player with nothing to switch on.

The export also hands you two caption files beside the video, an SRT and a VTT, for the places that want captions as a separate file. Each caption line in the editor becomes one cue in the files, with the same words and the same timing.

The export is the one step that uses credits, 20 credits a second with a minimum of 60 credits, and the cost is shown before you press it. Reading the words, styling them and previewing the result are all free to do.

Most captions are read on a phone held upright, so a caption has to be short enough to read at a glance. You set how many words show at once, from one up to six, and long lines wrap to stay inside the frame, so a sentence is shown a few words at a time instead of filling the screen.

For a tall video, pick the upright frame, put the captions in the lower third, and keep one or two words a page if the clip moves fast. For a wide video, show a full line at a time. The preview shows the result at the real frame, so you can judge it on the shape it will be watched in.