Guide

How to turn a podcast into 10 shorts

A lively one-hour podcast can hold ten or so moments that could work as shorts, and a slow one only a few. Here is how to find them, how to format them, and how to get them out without cutting each one by hand.

Last checked: 14 November 2026

What makes a good podcast clip

Most of a podcast is not clip material. People warm up, wander off and repeat themselves. The parts that work as shorts tend to share three things.

A clear claim or opinion in the first sentence. Viewers decide very quickly whether to keep watching. A clip that opens with "So, anyway, as I was saying" loses them. A clip that opens with a strong statement gives them a reason to stay.
One idea per clip, 20 to 60 seconds long. A short should make one point. If it needs the five minutes before it to make sense, it is not a clip yet.
An ending that lands. A punchline, a conclusion or a question. Stopping in the middle of a thought feels like a mistake, even when the middle was good.

When you listen back, mark the moments where someone tells a short story, disagrees, gives a number, or makes the room laugh. Those are your candidates. Skip sponsor reads, housekeeping and long setups.

Format

A good moment still needs the right shape for TikTok, YouTube Shorts and Instagram Reels.

9:16 vertical with the speaker's face centred. Most podcasts are filmed wide. Crop to the person who is talking, not to the middle of the table.
Large captions, 2 to 4 words at a time. Many people watch with the sound off. Short bursts of big text are easier to read than full sentences in small type.
Hook text on screen for the first 2 seconds. One line that says what the clip is about. You can add it in the platform's own editor when you upload.

Keep the important things away from the edges of the frame. Each platform covers part of the screen with its own buttons and text.

The less manual way: let clip-engine do the cutting

Doing all of this by hand in a video editor works, but finding, cropping and captioning ten clips takes hours. clip-engine is a Windows app that does the repetitive part on your own PC.

Load the episode. Open the video file from your computer (mp4, mov, mkv or webm). The video stays on your PC and is not uploaded.
It finds candidate moments. The app turns the speech into text, then looks for 20 to 60 second stretches that start and end on a sentence. By default it makes up to 10 clips per video.
It reframes them to vertical. Each clip comes out at 1080x1920. The app looks for faces and tries to follow the person whose mouth is moving. It can pick the wrong person, so check the framing when you review. Two people can be stacked one above the other. If it finds no face, it shows the whole frame over a blurred copy.
It adds captions. Captions are timed to each word. You can choose from several styles, and most of them highlight the word being spoken.
You review, then export the batch. Approve or reject each clip, fix caption words, edit the title and hashtags, and trim the start and end. Export gives you one MP4 per clip, plus a text file with its title, description and hashtags.

Keep the best ones and post one or two a day. That way one episode can fill a week of posts.

There are no credits or minute limits, so a long episode costs the same as a short one. More on that in a clipping tool with no credits.

Keep in mind: the picks are a starting point, not a final cut.

The app cannot tell a dull story from a gripping one. It scores things like sentence starts, reactions and laughter. You still choose which clips are worth posting, and some episodes will give you fewer than 10 good ones.
Speech-to-text mishears words. Read the captions on every clip before you export it.
Speed depends on your PC. In our own test on a 4-core laptop, a 59-minute video took about 27 minutes. That test video had no faces in it, so a real podcast may take longer. A graphics card does not make it faster.
You post the clips yourself. clip-engine does not upload or schedule anything. It can remind you when it is time to post while the app is open.
Windows only, and some internet is needed. It is built for Windows 10 and 11. So far it has been tested on Windows 11. The app goes online to activate, to download the speech model and a few small face and sound models once, to check your license and to look for updates. If you add your own AI key, it also sends the transcript text (never the video) to that AI service.

Clipping someone else's podcast

Only clip footage you own or have permission to use. If the podcast is not yours, get the owner's permission first. No tool changes who owns a video.

Some podcasts and brands invite clippers through Whop campaigns and provide the footage. If that is your plan, read how to start clipping on Whop. For choosing a tool for campaign work, see the best clipping tool for Whop. Campaign rules and rates differ, and many clips earn little or nothing. clip-engine is not made, endorsed or sponsored by Whop.

Questions people ask

How many shorts can I get from one podcast episode?

clip-engine makes up to 10 clips per video by default. How many of them are worth posting depends on the episode. A lively one-hour conversation may give you 10 good moments. A slow one may give you three or four.

How long does it take?

It depends on your PC. In our own test on a 4-core laptop, a 59-minute video took about 27 minutes from start to finished clips. That test video had no faces in it, so a real podcast may take longer. Your PC will be busy while it runs.

Can I use an audio-only podcast?

Yes. clip-engine opens mp3, m4a, aac and wav files as well as video. Clips made from sound-only files get a plain still background instead of video, not a moving waveform.

Does clip-engine post the shorts for me?

No. It exports each clip as an MP4 file with a text file holding the title, description and hashtags. You upload each one yourself to TikTok, YouTube Shorts or Instagram Reels.