Formats · Length 30-90s · 9:16 · beginner
Podcast clip video: how to shoot, edit, and prompt it
A podcast clip is a 30-90 second moment cut from a longer conversation, reframed vertical, captioned, and opened on the best line.
By Sulan Zhang, co-founder of Vyra · Updated 2026-09-24
- Platforms
- TikTok, Reels, Shorts, YouTube
- Length
- 30-90s
- Aspect
- 9:16
- Difficulty
- beginner
- Gear
- the podcast recording, one camera per speaker
The hard part is finding the moment in 90 minutes of talking. That's the part I let Vyra do.
At a glance
| Best for | Growing a podcast, guest promotion, reusing long interviews |
| Hook style | The punchline first, then the setup |
| Pacing | Switch speakers on every reply, punch in on emphasis |
| Caption style | Word-by-word with a color per speaker, or 3-5 word phrases |
| Music | None |
| Shot list | Wide two-shot. A single on each speaker. The episode title for the end card. |
| Common mistakes | Starting on the question instead of the answer. Clips that need context. One wide shot for 60 seconds. |
Structure
- 0-3s: the quotable line, big caption
- 3-10s: the question that led there, if needed
- 10-60s: the full answer, cut to whoever is talking
- 60-70s: the host's reaction
- 70-75s: end card with the episode name
How to shoot it
- One 4K camera per speaker so you can crop to vertical.
- Center each person with headroom so the 9:16 crop keeps the whole face.
- Separate audio tracks per speaker.
- Write down timestamps of good moments while recording. Saves an hour later.
- Same lighting on both people.
How to edit it
- Find a line that works with zero context. Start there.
- Reframe vertical on the active speaker, switching every turn.
- Cut filler and cross-talk hard.
- Captions with a small color change per speaker.
- Punch in 10-15% on the strongest sentence.
- End card, 3 seconds, episode title and where to listen.
Prompt it in Vyra
Find the 5 most quotable 45-60 second moments in this 90-minute podcast, the ones that make sense without context. Cut each one vertical, framed on whoever is speaking, with captions colored by speaker. Start each clip on the strongest line.
Cut the section where we talk about "burnout" into a 60-second Reel. Open on the guest's line about "the calendar was full but I was empty", then cut back to the question. Punch in on that line.
Match the layout and caption look of the attached reference clip for all five clips.
Variations
- Guest only, host cut entirely.
- Stacked layout, both speakers visible top and bottom.
- Audiogram, a photo with a waveform for audio-only shows.
FAQ
How do I pick which moment? Lines that are surprising, specific, or emotional and stand alone. If it needs the last 5 minutes, skip it.
Can Vyra find the moments? Yes. It transcribes the whole episode, so you can ask by topic, by keyword, or for "the most quotable moments".
Music or no music? No. The conversation is the sound.