YouTube Shorts Clip Finder - Best Moments to Repurpose

Paste a long YouTube video and get back the moments worth cutting into Shorts, Reels or TikToks — each with exact start and end seconds, the opening line as it was actually spoken, a title, a caption, hashtags and a score.

It reads the video's transcript with timestamps, so every pick is a real moment in the video, not a guess from the description.

What you get

One row per clip, best first:

• startSeconds / endSeconds and startTimestamp / endTimestamp — the exact cut points
• clipDurationSeconds — always inside the length window you set
• hook — the opening words, copied verbatim from the transcript
• title, caption, hashtags — ready to post, shaped for your target platform
• whyItWorks — one line on why the moment holds attention
• viralityScore — 1-100 short-form potential
• transcriptExcerpt — what is actually said inside the clip window, so you can sanity-check the pick before editing
• watchUrl — opens the video at the clip's start time
• clipRange — {url, start, end, label}, ready to paste into a clip-cutting actor
• Video context: videoTitle, channelTitle, uploadDate, videoDurationSeconds, viewCount

Use cases

• Repurposing a back catalogue — mine podcasts, interviews and lectures for short-form content
• Podcast clip production — hand an editor exact in and out points instead of "somewhere around the middle"
• Agency workflows — process a client's long videos in bulk and deliver a ranked clip sheet
• Competitor teardown — see which moments of a rival's video would travel as Shorts
• Newsletter and social scheduling — turn one upload into a week of posts

How to use

1. Enter one or more Video URLs (long-form videos with captions work best).
2. Set Clips per video and the Minimum / Maximum clip length.
3. Pick the Target platform so captions and hashtags match where you are posting.
4. Run the actor — clips appear in the Dataset tab, ranked.
5. Take startSeconds / endSeconds into your editor, or paste clipRange into a clip-cutting actor to get the file itself.

Output format

{
  "rank": 1,
  "startSeconds": 439,
  "endSeconds": 494,
  "startTimestamp": "7:19",
  "endTimestamp": "8:14",
  "clipDurationSeconds": 55,
  "title": "The shocking truth about drug efficacy",
  "hook": "It's for the number needed to treat. Now, what is this, the NNT?",
  "caption": "Most drugs work for far fewer people than you think. Here's the number nobody quotes.",
  "hashtags": ["#MedicalTruth", "#NNT", "#Healthcare"],
  "whyItWorks": "Opens on a counterintuitive statistic and pays it off inside a minute.",
  "viralityScore": 95,
  "watchUrl": "https://www.youtube.com/watch?v=VIDEO_ID&t=439s",
  "clipRange": { "url": "https://www.youtube.com/watch?v=VIDEO_ID", "start": "7:19", "end": "8:14", "label": "the-shocking-truth-about-drug-efficacy" }
}

Notes

• The video needs captions — manual or auto-generated. Videos with captions disabled come back with an explanation in the error field.
• Clips never overlap: a candidate that covers the same ground as a higher-ranked pick is dropped.
• Timestamps come back as numbers, so they drop straight into an editor, ffmpeg, or a clip-cutting actor without reformatting.
• Longer videos cost a little more to analyse but are where this pays off — a two-hour podcast usually yields several strong clips.