The three sources of B-roll
Your own footage
Your own footage
The most authentic option. Screen recordings, product demos, behind-the-scenes clips, photos from your camera roll: anything that visually supports what youâre saying.Best for: personal brands, product demos, tutorials, any content where showing your specific context matters.
Stock imagery and video
Stock imagery and video
Captions has a built-in stock library with photos, videos, GIFs, and stickers. Searchable and free to use.Best for: general concepts, supplementary visuals, quick fills when you donât have specific footage.
AI-generated images and video
AI-generated images and video
Type a prompt and generate a custom image or short video clip. Powered by multiple AI models including Flux, Imagen, DALL-E, Pika, Ray, and more.Best for: original, on-brand visuals; concepts that are hard to find in stock libraries; maximum creative control.
Adding B-roll manually
Choose your source
- Import: select from your camera roll or files
- Import â Stock: search the stock library for photos, videos, GIFs, or stickers
- Generate: enter a prompt to generate a custom image or video clip
Place and resize
After adding the media, it appears as an overlay in the timeline. Drag it to position it over the right moment. Pinch to resize. Move it on the canvas by dragging with one finger.
Generating B-roll with AI
Choose image or video
Select whether you want a still image or a short video clip. Video clips are more engaging but take longer to generate.
Write your prompt
Describe what you want. Be specific:
- â âa person using a phoneâ
- â âclose-up of hands typing on a smartphone screen, warm natural lighting, shallow depth of fieldâ
Choose your AI model
Different models have different strengths. For photorealistic images: Flux 1.1 Pro, Google Imagen 3, or DALL-E 3. For stylized or illustrated looks: Ideogram. For video: Pika 2.2 or Luma Ray 2.
Editing B-roll after placing it
Tap any media overlay in the timeline to edit it:- Position presets: Caption Aware (auto-repositions so B-roll doesnât cover captions), Top Half, Bottom Half, Full Screen, Freeform
- Transitions: Fade, Pop, Zoom, Scale, Slide Left, Slide Up, or None
- Animate: for still images, tap Animate to add slow zoom or pan motion
- Replace: swap the media with a different image or clip
- Split: split the media overlay at the playhead
Pro tips
- Use Caption Aware positioning. It automatically repositions B-roll so it doesnât cover your captions, which saves time and looks more polished.
- Animate still images. A slow Ken Burns zoom on a photo looks significantly more professional than a static image just sitting on screen.
- Match the B-roll to the speech. B-roll works best when it visually illustrates exactly whatâs being said at that moment, not just something vaguely related.
- Keep clips short. 2â5 seconds per B-roll segment keeps the edit moving. Longer clips need to be earning their place.
- Use transitions sparingly. Fade works everywhere. Pop and Zoom can be effective. Donât use a different transition on every clip.
Media Overlay reference
Full feature documentation.
Go from Raw Footage to Finished Video
Let AI Edit place B-roll automatically.

