Skip to content
promptbeginnerFree

Podcast clip B-roll video prompt

Turn a podcast quote into 3 short B-roll cutaways that illustrate what's being said, so clips aren't just two people at microphones.

You'll need: A transcript excerpt from your podcast · A text-to-video model

Fill in your details

Values are inserted into every prompt below. Nothing you type leaves your browser.

4 field(s) left — unfilled variables stay as {{placeholders}}.

  1. 1Generate B-roll prompts from the quote

    Run in a chat model. Cut the B-roll over the speaker's audio for 2-3 seconds at a time, then return to the speaker.

    Here is a 30-60 second excerpt from my podcast:
    """{{transcript}}"""
    Podcast topic: {{topic}}. Visual style for all cutaways: {{visual_style}}. Aspect ratio: {{aspect_ratio}}.
    
    1. Pick the 3 phrases that most benefit from a visual (concrete images, metaphors or numbers).
    2. For each, write a video prompt for a 3-second cutaway: [shot and camera move]. [literal or metaphorical visual]. Lighting: [...]. Style: {{visual_style}}. Duration 3 seconds, aspect ratio {{aspect_ratio}}. Avoid: text, logos, real people, warped faces.
    3. Say where in the excerpt each cutaway should start.

Example input

Excerpt: 'Most freelancers treat their bank account like a bucket with a hole in it...' Topic: freelance finance; 9:16.

Example output

Cutaway 1 at "a bucket with a hole in it" Close-up, static camera. A metal bucket on a wooden floor slowly leaks a stream of coins from a hole in its side. Lighting: warm window light from the left. Style: warm cinematic, soft natural light, shallow depth of field. Duration 3 seconds, aspect ratio 9:16. Avoid: text, logos, real people, warped faces.

Troubleshooting

  • Visuals are too literal and cheesy: ask for 'subtle, everyday visual metaphors'.
  • Clips clash with the studio footage: match the grade (warm vs cool) to your camera footage.

Known limitations

  • Cutaways can change how a quote is perceived; don't use imagery that implies something the speaker didn't say.

More from Video generation