🚀Publish FREEJoin WriterDock's early community of writers.Start Writing — It's Free

AI Clipper: How to Turn One Long Video into 3 Viral Shorts in Under a Minute

Suraj Kumar

Suraj Kumar

Founding Writer #1Contributor

November 21, 2026•9 min read•7 views

AI Clipper: How to Turn One Long Video into 3 Viral Shorts in Under a Minute

GTA 6 hype is everywhere, and creators are about to upload three-hour gameplay sessions by the dozen. But there is a problem nobody talks about: nobody watches a three-hour video from start to finish. The real growth is in Shorts, Reels and TikToks, where a single strong 40-second moment can reach more people than the full video ever will.

The catch is that finding that moment is slow. You scrub through the timeline, guess which part is interesting, cut it, crop it to vertical, add captions, and export. For a 10-minute video, that easily eats an hour or two. Do it for every upload and you have a second full-time job.

This is where an AI Clipper comes in. In this article, I'll break down the exact three-step workflow shown in my Reel: how it takes a 10-minute video and turns it into three ready-to-post Shorts of about 40 seconds each. At the end of every section you'll find the download links for the tools and files you need.

What Is an AI Clipper?

An AI Clipper is a simple pipeline that does the boring editing work for you. Instead of you watching the whole video to find the good parts, software listens to the video, a language model decides which moments are strongest, and a video tool cuts and reformats them automatically.

The workflow uses three building blocks:

  1. Whisper turns the video's audio into a timestamped transcript.
  2. Claude reads that transcript and picks the best hook moments.
  3. FFmpeg cuts those moments out and converts them from horizontal (16:9) to vertical (9:16).

The result: a 10-minute video becomes 3 Shorts, about 40 seconds each, and the whole thing runs in roughly fifty seconds of processing time.

It works for almost any long-form content. Gaming footage, podcasts, lectures, class recordings, interviews, tutorials. If someone is talking or something interesting is happening, it can be clipped.

Step 1: Create a Transcript with Whisper

Everything starts with text. A video is hard for a computer to "skim," but a transcript is easy.

Whisper is a speech-to-text model that listens to your video's audio and produces a transcript with a timestamp on every line. For example:

  • 02:44, "Want to get out of town instead?"
  • 02:51, "Sounds like an adventure."
  • 03:22, "...they owed the company money."
  • 04:06, "...the Danbury shakes."
  • 07:51, "Radium."

Those timestamps are the key. Once every sentence has a time attached to it, the rest of the pipeline knows exactly where to cut.

A few practical tips for this step:

  • Use clean audio. Whisper is accurate, but background music or heavy noise lowers the quality of the transcript.
  • Pick the right model size. Smaller models are faster; larger models are more accurate. For most creator content, a medium-sized model is a good balance.
  • Keep the timestamps. Don't strip them out when you save the transcript. You'll need them in Step 3.

Download links for this step:

  • Whisper download / install guide: click here
  • Python install (required to run Whisper): click Here

Step 2: Let Claude Pick the Hooks

This is the step that makes the whole workflow smart instead of random.

You hand Claude the full transcript and ask it to find the moments most likely to hold a viewer's attention. Claude reads the entire thing, which would take a human ten minutes or more, and returns the strongest hooks along with their timestamps.

In the example from the Reel, a video about the worst jobs in American history produced three clear hooks:

  1. "Want to get out of town?" followed by "...owed the company money." A curiosity-driven opening with a surprising payoff.
  2. "Hat makers had a problem." followed by "Mad as a hatter... was real." A familiar phrase that suddenly gets an explanation.
  3. "The secret ingredient? Radium." followed by "50 dial workers had died." A shocking reveal with real stakes.

Notice the pattern. Every good clip has two parts: a hook that makes people stop scrolling, and a payoff that rewards them for staying. A clip with a great hook and no payoff feels like clickbait. A clip with a payoff and no hook never gets watched. Claude is useful here because it can find both pieces and make sure they sit close enough together to fit in a 30 to 60 second Short.

How to get the best results from this step:

  • Ask for a specific number of clips (for example, three) so you get focused picks instead of a long list.
  • Give a target length, such as 30 to 45 seconds.
  • Ask it to return the start time, end time, and the hook line for each clip.
  • Tell it what kind of content it is, such as gaming, podcast or lecture, so the picks match your audience.

Download links for this step:

Step 3: Cut and Convert to Vertical with FFmpeg

Now you have a list of timestamps. The last step is turning them into actual video files.

FFmpeg is a free, open-source command-line tool that can cut, crop, resize and convert almost any video format. It is the engine behind a huge number of video apps, and it is perfect for this job because it can do the work at every exact moment the transcript says, without you opening an editor.

For each clip, FFmpeg does two things:

  1. Cuts the video from the start time to the end time Claude chose.
  2. Converts 16:9 to 9:16, reframing the horizontal video into a vertical one that fills a phone screen.

In the Reel, you can see this happen. A wide, landscape video becomes a tall vertical clip, with the subject centered and captions placed near the bottom of the frame. Captions matter more than most people think, because a large share of Shorts and Reels are watched without sound.

Some things to keep in mind:

  • Check the crop. A center crop works for many videos, but if the important action is on the left or right, you may need to shift the crop window.
  • Add captions. The same Whisper transcript you made in Step 1 can be turned into subtitles, so you already have the text.
  • Export at a platform-friendly resolution. A vertical 1080 x 1920 export is the standard for Shorts, Reels and TikTok.

Download links for this step:

  • FFmpeg download (Windows / Mac / Linux): Click Here

The Full Result: 3 Shorts in About 40 Seconds Each

Put the three steps together and the workflow looks like this:

StepToolWhat it does
1WhisperConverts audio to a timestamped transcript
2ClaudeReads the transcript and picks the best hooks
3FFmpegCuts each moment and converts it to vertical

A 10-minute video goes in. Three vertical Shorts, each around 40 seconds, come out, in about fifty seconds of processing. What used to take a couple of hours of manual scrubbing and exporting now happens while you make a cup of chai.

And because the pipeline only needs a video with speech or a clear storyline, it works well for:

  • Gaming. Long sessions with funny moments, big wins, or reactions.
  • Podcasts. Strong opinions and quotable lines are easy to find in a transcript.
  • Lectures and classes. Turn one long class into several bite-sized lessons.
  • Interviews and webinars. Pull out the single most useful answer.

Who Should Use This?

If you publish any long-form video and want to grow on short-form platforms, this workflow is worth trying. It is especially useful if you are:

  • A gaming creator who records long sessions and has no time to edit.
  • A podcaster who wants to promote episodes with short teasers.
  • A teacher or coach who wants to repurpose lectures into short lessons.
  • A small business or agency that needs a steady stream of short content without hiring an editor.

A quick honest note: AI picks are a strong first draft, not a final decision. Always watch the three clips before you post. Sometimes a hook is good but the clip ends one sentence too early, and a quick trim fixes it. Think of the AI as a fast assistant that does 90 percent of the work, and you as the editor who makes the final call.

Get the Complete Tool

Everything above is how the AI Clipper works under the hood. If you want the complete ready-to-use tool, without setting up each piece yourself, here are the links:

Frequently Asked Questions

Do I need to know coding to use this?

Basic comfort with copying and pasting commands is enough. The setup guide above walks through each step, and the ready-made scripts handle the rest.

Is it free?

Whisper and FFmpeg are free and open source. Claude has a free plan, and paid plans are available if you need more usage.

Can it work with Hindi or Hinglish videos?

Whisper supports many languages, including Hindi. Accuracy is usually best with clear audio, so test it on one short video first.

How long does it take?

In the Reel's example, a 10-minute video was processed in around fifty seconds. Longer videos take proportionally longer, and speed depends on your computer.

Will the clips be ready to post immediately?

Often yes, but always review them first. A small trim or caption fix can make a good clip better.

Final Thoughts

Short-form video rewards speed and volume, and most creators lose the race before they start because editing eats all their time. An AI Clipper flips that. Whisper listens, Claude finds the hooks, and FFmpeg cuts and reformats, so you can turn one long upload into several Shorts in the time it takes to read this sentence.

Download the resources above, try it on your next video, and see how many Shorts you can get out of a single recording. If you found this useful, follow for more creator tools and workflows.

About the Author

Suraj Kumar

Suraj Kumar

@surajkumar

Founding Writer #1Contributor

Suraj Kumar is a writer, entrepreneur, and the CEO and founder of this website, sharing simple and practical insights on business, creativity, and personal growth. With experience building digital projects, they enjoy helping others learn, grow, and succeed online.

Read Next

More articles you might enjoy

View all articles
7