Skip to content

Marketing & Content

Video Captions & Snippet Clips

Long videos become captioned clips and captions, mapped to timestamps.

Saves about ~3 hrs per week

What it does

  • Transcribes new videos automatically
  • Finds the best 3 moments for short clips with timestamps
  • Generates clean captions for every clip
  • Outputs a shot list you can hand to any editor

How it works

  1. Video dropped into the watch folder
  2. Transcription runs automatically
  3. Best moments identified with timestamps
  4. Captions and shot list generated

What's included

  • Auto-transcription
  • Best-moment selection
  • Caption output

Set it up yourself

Follow these 4 steps in order. You will need accounts for: Claude.

Step 1: Set up the watch folder

  1. Create a Drive folder 'Videos to Clip' and connect it to your workflow as the trigger.
  2. Decide the video sources: uploads there, or a YouTube channel trigger.
  3. For YouTube: connect your channel so new uploads auto-enter the flow.
  4. Name files by episode/date so outputs map cleanly.
  5. Test an upload triggers the flow.

Watch out: Drive triggers fire per folder — do not watch a folder that contains already-clipped videos or you re-process everything.

Step 2: Add the transcription step

  1. Connect a transcription service: Whisper API, or the one your video tool exports (Descript, Otter).
  2. If using Whisper: send the audio, get back segments with timestamps.
  3. Store segments as JSON: {start, end, text}.
  4. Test on a 10-minute video and check segment quality.
  5. For long videos chunk the audio (Whisper handles ~5min chunks best).

Watch out: Timestamps are the whole point — keep the segment timestamps intact, losing them destroys the shot list.

Step 3: Add the clip selection prompt

  1. Build: 'From these transcript segments pick the 3 best 30-60s moments for short vertical clips. Priority: standalone punchlines, specific numbers, self-contained stories. Return JSON with start, end, reason, suggested caption.'
  2. Instruct: moments must make sense out of context (no mid-story cuts).
  3. Ask for clean caption text (add punctuation, remove filler words, marked).
  4. Paste with API key.
  5. Test and eyeball the edges.

Watch out: 'Makes sense standalone' is the hardest constraint — a clip that needs the prior minute is a dead clip.

Step 4: Export the shot list

  1. Write the output to a sheet 'Clip Shotlist': Video, Clip #, Start, End, Reason, Caption, Status.
  2. Hand the sheet to your editor or use it as the brief for a clipping tool.
  3. For auto-editing: many tools accept the timestamps directly.
  4. Mark Status when each clip ships.
  5. Review quarterly whether the selects are worth a caption pass.

Watch out: The shot list is the deliverable — update it as clips publish so you don't re-select the same moments twice.

Frequently asked questions

What does the Video Captions & Snippet Clips automation do?

Long videos become captioned clips and captions, mapped to timestamps. Transcribes new videos automatically

How much time does Video Captions & Snippet Clips save?

It is estimated to save ~3 hrs per week. Actual savings depend on your volume.

Which tools does Video Captions & Snippet Clips use?

It works with Claude.

What does Video Captions & Snippet Clips include?

Auto-transcription, Best-moment selection, Caption output.

Can I set up Video Captions & Snippet Clips myself?

Yes. The do-it-yourself guide on this page walks through 4 steps, from getting access to testing the finished automation.