Put your own b-roll over the line it belongs to

How placement works

A photo, or a video with no speech in it, is treated as b-roll. Cutroom AI compares the words in its file name with the words you speak and places it over the line that matches. Name a photo garlic-closeup.jpg and it lands over "Then slice the garlic thin." If a b-roll clip has some speech in it, those words count too.

Each line gets at most one piece of b-roll, so two photos named after the same line don't stack. Photos fill the width of the frame and drift in, so they don't sit there like a slide.

A real case: a sourdough tutorial with three takes of you talking at the counter, plus starter.jpg, folding.mov and crumb.jpg. The edit opens on your best take of "Feed the starter the night before" with starter.jpg drifting in over it, cuts to the folding clip on "keep folding until it holds", and ends on the crumb shot when you say "look at that crumb".

Name your files before you import

File names are the whole signal, so a minute of renaming pays off. Patterns that work:

• Use a word you actually say: kitchen.jpg for "the kitchen was redone last year".
• Use the noun, not the event: sourdough-crumb.mov rather than saturday-bake.mov.
• Drop camera names. IMG_2042.jpg and DSC0193.mov match nothing.
• One idea per file. A photo can only go on one line, so pool.jpg and firepit.jpg place better than one file named after both.

Where unmatched photos go

Photos whose names didn't match a line wait in the Overlays tab under "Photos to place". Move the playhead to where one belongs and press "Add at playhead". You can change the start time, length, position and size of any placed piece there. An edit holds up to 20 overlay layers in total, counting text, photos, video and sound.

How to use it

1. 01Rename your photos and silent clips after the words you say over them.
2. 02Open the editor, choose Talking head and drop in your takes and the b-roll files together.
3. 03After the analysis, play the edit. Each matched file sits over its line.
4. 04Open the Overlays tab to place anything left in "Photos to place" and to adjust timing.
5. 05Export as usual. The b-roll is rendered into the MP4 on your device.

What it can't do

There's no image recognition. Cutroom AI never looks at what's in the picture, so it can't tell a photo of a kitchen from a photo of a cat. If the name says nothing, the file waits for you.

There's no stock footage and no generated b-roll. Everything on screen is something you filmed or photographed.

Photos only go over talking-head edits. In montage mode, only video clips are cut to the beat.

When another approach is better

If you need shots you don't have, say a city skyline for a market update, get them from a stock library or a friend, rename them after your line and import them like your own. They're placed the same way.

If you have lots of b-roll and very little talking, a montage cut to a song may suit the footage better than a talking-head edit.