Demo · Claude

Edit a two-person interview into a Short

A podcast host asks Claude to cut six camera angles of an interview together. cuttingfloor cuts to whoever is talking, removes the dead air and the ums, adds name cards and captions, and makes vertical and widescreen versions.

01

Before

The clip as it came in · 0:48

interview-all-angles.mp4 · 0:48

What needs to go

[2 s before Marcus starts] Okay, so you went from zero to a hundred thousand followers in one year. [pause] How? [silence while the camera changes] Honestly? I posted every single day, even when, um, nobody was watching. (Marcus laughs) Every single day? Hmm. Every day. The first fifty. But that’s how you get good. [3 s of silence] So what’s your one tip for someone starting today? Um… [3 s of silence] Start before you’re ready. Just post the thing.

Struck in red: 8 things the agent cuts.

02

The edit

In Claude, with cuttingfloor connected

[2 s before Marcus starts] Okay, so you went from zero to a hundred thousand followers in one year. [pause] How? [silence while the camera changes] Honestly? I posted every single day, even when, um, nobody was watching. (Marcus laughs) Every single day? Hmm. Every day. The first fifty. But that’s how you get good. [3 s of silence] So what’s your one tip for someone starting today? Um… [3 s of silence] Start before you’re ready. Just post the thing.

interview-short.mp4 · 0:23
Claudecuttingfloor
  1. Cut this interview together. Show whoever’s talking and tighten the gaps.

    I’ll read every angle, find exactly where each person talks, then cut to whoever’s speaking and close up the gaps so it plays like one conversation.

    • Used cuttingfloor › skills.get read the cleanup steps
    • Used cuttingfloor › media.import six angles
    • Used cuttingfloor › media.analyze every word, timed
    • Used cuttingfloor › code.run measured where each person talks
    • Used cuttingfloor › render.frames looked at each angle
    • Used cuttingfloor › edit.batch 7 shots, cut to the speaker
    • Used cuttingfloor › edit.batch sound leads into each shot
    • Used cuttingfloor › audio.set_master louder, even
    • Used cuttingfloor › render.frames checked every cut
    • Used cuttingfloor › export.video widescreen, 0:23

    Done: 48 seconds of takes down to 23. The camera follows whoever’s speaking, and the um is gone. Want names and captions?

  2. Add their names and captions, and make me a vertical one for Shorts.

    I’ll put a name card on each person the first time we see them, caption every line in that person’s colour, then make a vertical version that keeps whoever’s talking in frame.

    • Used cuttingfloor › code.run drew the name cards
    • Used cuttingfloor › media.analyze found the faces
    • Used cuttingfloor › caption.add every line, per speaker
    • Used cuttingfloor › text.set gold for him, pink for her
    • Used cuttingfloor › edit.batch placed the name cards
    • Used cuttingfloor › code.run vertical, framed on faces
    • Used cuttingfloor › render.frames checked the vertical
    • Used cuttingfloor › export.video vertical and widescreen

    Done: name cards for Marcus and Dana, captions in gold for him and pink for her, and a vertical Short framed on whoever’s talking. The widescreen one is ready too. Want a hook title at the start?

Shown in your agent: Claude, ChatGPT or Kazi. cuttingfloor does the editing and hands back the file; your agent shows it to you.

03

After

Ready to post · 0:48 → 0:23

interview-short.mp4 · 0:23

What the agent did

  1. Read all six camera angles and wrote down every word.
  2. Measured where each person actually speaks, so the cuts land on speech.
  3. Found the guest’s “um”, which the transcript left out.
  4. Cut to whoever is talking and trimmed the dead air around every line.
  5. Let each line’s sound start just before its shot, so the cuts flow.
  6. Hid the um cut under a shot of Marcus listening, so nothing jumps.
  7. Made the sound louder and even.
  8. Added a name card for each person and captions in their own colour.
  9. Made a vertical version that follows each speaker’s face, and a widescreen one for YouTube.
Widescreen, for YouTube interview-widescreen.mp4 · 0:23