YouTube Most Replayed Moments Scraper

Find the exact moments viewers rewatch most in any YouTube video or Short. This actor reads YouTube's "Most replayed" heatmap, the replay graph shown above the progress bar, and turns it into a ranked list of peak moments. Each moment comes with start and end timestamps, a replay intensity score, a deep link that opens the video at that second and, if you want it, the words spoken during that moment. No login, no cookies, no YouTube API quota.

Paste video URLs and export the results to JSON, CSV, Excel or Google Sheets.

What you get

One row per video:

• videoId, url, title, channelTitle, channelId, channelUrl
• durationSeconds, duration ("57:51"), views, isShort
• heatmapAvailable: whether YouTube shows a "Most replayed" graph for this video
• mostReplayed: the moment YouTube itself labels "Most replayed", with timestamp and deep link
• peaks: your top replayed moments, ranked by intensity. Each peak has:
  • rank, startSeconds, endSeconds, peakSecond
  • intensity: replay intensity from 0 to 1 (1 = the most replayed point of the video)
  • timestamp ("46:16") and peakTimestamp ("46:34")
  • url: a deep link (&t=2776s) that opens the video at that moment
  • isYouTubeMostReplayed: true when YouTube's own label falls inside this peak
  • isIntro: true for the opening-second spike
  • transcript: what is said during the moment (optional)
• heatmap: the full replay curve, about 100 points with start time, length and intensity (optional)
• transcriptLanguage, transcriptIsAutoGenerated (when transcripts are on)

Use cases

• Clip hunting: find the moments worth cutting into Shorts, Reels and TikToks from podcasts, streams and long videos
• Content research: see which part of a competitor's video viewers actually rewatch: the reveal, the joke, the tutorial step
• Video editing and retention analysis: compare replay peaks across your own uploads to learn what keeps people watching
• Podcast highlights: pull the most replayed segments of long interviews together with the quote that made them replayable
• Data and dashboards: export the raw heatmap curve to chart audience attention across many videos

How to use

1. Add one or more Videos: watch URLs, youtu.be links, Shorts URLs or bare video IDs.
2. Set Top peaks per video (default 5).
3. Set Minimum gap between peaks: peaks closer than this are merged into one moment. 20 seconds suits most videos, while 60 to 180 seconds works better for long podcasts.
4. Turn on Include transcript for each peak if you want the spoken words for every moment.
5. Leave Include full heatmap on if you want the raw curve for your own charts.
6. Run the actor. Results appear in the Dataset tab.

Which videos have a heatmap?

YouTube only shows the "Most replayed" graph once a video has enough repeat viewing, so most small or very new videos do not have one. In our testing even videos with 100M+ views had no graph during roughly their first two weeks. These videos are still returned, with heatmapAvailable: false and empty peaks, and you are not charged for them.

Shorts usually have a curve that starts high and fades out, because every loop starts from the beginning. When a video has no peak besides its opening second, that intro moment is returned as the single peak (isIntro: true).

Output format

{
  "videoId": "Af6i6ChAVTw",
  "url": "https://www.youtube.com/watch?v=Af6i6ChAVTw",
  "title": "Last To Leave Mansion, Keeps It",
  "channelTitle": "MrBeast",
  "channelId": "UCX6OQ3DkcsbYNE6H8uQQuVA",
  "channelUrl": "https://www.youtube.com/@MrBeast",
  "durationSeconds": 3471,
  "duration": "57:51",
  "views": 131616463,
  "isShort": false,
  "heatmapAvailable": true,
  "mostReplayed": {
    "label": "Most replayed",
    "second": 2534,
    "timestamp": "42:14",
    "url": "https://www.youtube.com/watch?v=Af6i6ChAVTw&t=2534s"
  },
  "peaks": [
    {
      "rank": 1,
      "startSeconds": 2776,
      "endSeconds": 2812,
      "peakSecond": 2794,
      "intensity": 0.9,
      "timestamp": "46:16",
      "peakTimestamp": "46:34",
      "url": "https://www.youtube.com/watch?v=Af6i6ChAVTw&t=2776s",
      "isYouTubeMostReplayed": false,
      "isIntro": false,
      "transcript": "It's DONE! ... I have to win this mansion."
    }
  ],
  "heatmap": [
    { "startSeconds": 0, "durationSeconds": 34.71, "intensity": 1 },
    { "startSeconds": 34.71, "durationSeconds": 34.71, "intensity": 0.3465 }
  ],
  "status": "success",
  "error": null
}

How peaks are found

• Every local high point on the replay curve is a candidate. Small bumps in flat stretches are ignored.
• Candidates are ranked by intensity. A candidate is dropped if it sits within Minimum gap between peaks of a stronger peak or inside its window.
• Each peak's window grows outward while the curve stays at 70% or more of the peak's intensity, up to about 5% of the video on each side.
• The heatmap has about 100 points per video, so each point covers roughly 1% of the running time (about 2 seconds on a 3-minute video, about 2 minutes on a 3-hour podcast). Timestamps are as precise as YouTube's own graph.

Pricing

Pay per event:

• video-result: charged once for each video that has most-replayed data. Videos without a heatmap are free.
• transcript-result: charged once for each video that gets transcript text attached to its peaks (only when Include transcript is on and the video has captions).

FAQ

Can I get the most replayed part of a YouTube Short? Yes. Paste the Shorts URL. The heatmap and peaks work the same way.

Why is the top peak different from YouTube's "Most replayed" label? YouTube's label sometimes points to a moment other than the highest point on the curve, and this actor ignores the opening-second spike by default. Both are returned: mostReplayed is YouTube's label, and peaks is the ranked curve analysis.

Does it work for private, members-only or age-restricted videos? No. Only public videos are supported.