Plan and Add B-roll with AI
B-roll is the supporting footage that plays over your voice: a stock clip, a screen recording, a photo, a graphic. It keeps eyes on the screen, hides your cuts, and makes a talking-head video feel like a produced piece instead of a webcam monologue. Choosing what B-roll to add and where is the creative part, and that is exactly where an AI planning partner shines. This lesson uses AI to build a shot list, then adds the footage in CapCut.
What You'll Learn
- What B-roll is and why it transforms a talking-head video
- How to use AI to generate a B-roll shot list from your script
- Where to find free B-roll footage, including CapCut's built-in library
- How to layer B-roll in CapCut without covering your captions
Why B-roll Changes Everything
Watch any polished YouTube video or ad. The creator talks, but you are not staring at their face the whole time. You see the thing they mention: the app they demo, the city they describe, a chart of the number they cite. That is B-roll, and it does three jobs.
It holds attention. A changing image resets the viewer's focus. Static shots lose people.
It hides cuts. Remember the visual jumps from silence removal and structural cuts? Lay B-roll over them and the join vanishes, because the viewer is looking at the B-roll, not your jumping head.
It shows instead of tells. Saying "the interface is cluttered" is weaker than showing the cluttered interface while you say it.
You do not need much. Even three or four well-placed B-roll clips lift a short video noticeably.
Build a Shot List with AI
The hard question is "what B-roll and where?" Answer it with your transcript. Paste this into ChatGPT, Claude, or Gemini:
Here is the transcript of my [topic] video. Suggest specific B-roll for it. For each suggestion give: the exact line it should play over, what the shot should show, and whether it is best sourced as stock footage, a screen recording, a photo, or a simple graphic. Aim for [4-6] B-roll moments. Keep suggestions concrete and searchable.
The AI returns a shot list tied to your actual words: "Over 'editing used to take hours,' show a timelapse of someone editing at a desk. Source: stock." That is a plan you can execute. It turns a vague creative task into a checklist.
Ask a follow-up for search terms:
For each stock suggestion, give me three search phrases I could type into a free stock video site to find it.
Now you know exactly what to search for.
Where to Get Free B-roll
You have several free sources.
CapCut's built-in library. CapCut includes stock video, images, stickers, and graphics right in the media panel. Search a keyword, drag a clip to the timeline. This is the fastest source and covers most common needs. Check any licensing notes for commercial use.
Free stock sites. Pexels, Pixabay, and Mixkit offer free-to-use video clips. Download and import them into CapCut. Use the AI's search phrases to find matches. Always glance at the license, most are free even for commercial use, but confirm.
Your own footage. A screen recording of the app you are discussing, a photo you took, a quick clip on your phone. Original B-roll is unique and always properly licensed because it is yours.
Simple graphics. A title card, an arrow pointing at something, a number on screen. CapCut's text and sticker tools make these in seconds. For anything more designed, a tool like Canva exports an image or clip you import.
Use Perplexity if you need to find a source fast:
Free, no-attribution stock video sites I can use for a [topic] video, with a note on each one's license. List 5.
Layer B-roll in CapCut
B-roll goes on a track above your main clip so it plays on top.
- Move the playhead to where the B-roll should start.
- Drag your B-roll clip onto an upper track, above the main video.
- Trim its length by dragging its edges so it covers exactly the line it belongs with.
- The B-roll now hides your face; your voice keeps playing underneath because the audio is on the main clip below.
Two adjustments make it clean. Match the duration to the point so the B-roll leaves when the topic changes. And keep your captions on top of everything, on the highest track, so the B-roll never covers them. Check the layer order: captions above B-roll above main video.
If a B-roll clip has its own loud audio (music, ambient noise), mute it so it does not fight your voice. Select the clip and drop its volume to zero.
Do Not Overdo It
B-roll is seasoning, not the meal. Too much and the viewer never connects with you; the video feels like a stock-footage reel. Return to your face for the important, personal moments, and use B-roll to illustrate, cover cuts, and add energy. Well-placed beats constant.
Try It Now
- Run the B-roll shot-list prompt on your transcript.
- Pick three suggestions. Find footage in CapCut's library or a free stock site using the AI's search phrases.
- Layer each clip on a track above your main video, trimmed to its line, with any B-roll audio muted.
- Confirm your captions still sit on top and are never covered.
Key Takeaways
- B-roll holds attention, hides cuts, and shows instead of tells; even a few clips help.
- AI turns your transcript into a concrete shot list with sources and search terms.
- Free B-roll comes from CapCut's library, Pexels/Pixabay/Mixkit, your own recordings, and simple graphics.
- Layer B-roll on a track above the main clip, trim it to its line, mute its audio, and keep captions on top.
- Use B-roll to season the video, not to replace your presence.

