How to Turn a Video Podcast Into Shorts and TikToks

To turn a video podcast into Shorts or TikToks, you do five things: pick the passages that stand on their own, reframe them to vertical (9:16) without losing your guests, add captions, give each clip a title that holds people in the first seconds, and post at a pace you can actually keep up. This guide walks through each step, by hand and with a tool.

Full disclosure: we build Nysos, which handles a good part of this from a link or a video file. We'll say where it helps and where your own eyes matter. Nysos starts from a video: a YouTube, Twitch or Kick link, or a video file.

Which parts of a podcast make good clips?

A good podcast clip isn't a random excerpt. It's a passage that someone who has never heard the episode can follow from start to finish. Four shapes come up again and again.

• A complete idea. Someone sets out a point, explains it and lands it. If you can sum the passage up in one sentence, it's a strong candidate.
• A question and its answer. The question works as the hook and the answer as the substance. Keep the question as the host asked it: it supplies the context without anyone having to explain it.
• A story. An anecdote with a beginning, a build and a payoff. It cuts cleanly when the guest gets straight to the point instead of winding up for a minute.
• A disagreement. Two people who don't see it the same way: the tension comes through even without context, and it makes viewers stay to find out who's right.

Skip anything that leans on what was said twenty minutes earlier (“as we covered earlier”), plus transitions, thank-yous and sponsor reads. If you have to explain the setup for the clip to make sense, it's the wrong passage. A one-hour episode usually holds several decent passages and very few excellent ones: post the excellent ones. And if a guest shares something personal, check that they're comfortable with that moment travelling on its own.

How long should a podcast clip be?

As long as the idea needs. Our rule of thumb, the same one as in our clipping guide, runs from 30 seconds to 2 minutes: below that you rarely have time to tell anything, and beyond it the idea has to be very strong to keep someone to the end.

Two habits help you trim without damage: start on the first useful sentence rather than on “so, um, anyway”, and stop right after the payoff. In Nysos, each clip's page lets you move the start and end to the second and cut a section out of the middle, as long as the source is still stored (7 days after processing, 30 days with a subscription or for an imported file). Genuine silences (very low sound and no spoken word for at least 0.35 seconds) are removed automatically, except on a clip whose start or end you've adjusted by hand.

How do you frame two people in a vertical video?

This is the podcast headache: the source is landscape with two people side by side, and the clip is vertical. A center crop cuts one of them off, sometimes both. There are three ways out.

• One face in the shot: crop a vertical column that follows that face.
• A two-person conversation: a split screen stacks one face on top and one below, each taking half the height. Both people stay on screen, which keeps the exchange alive.
• No clean crop available (a face that's hard to track, a busy shot): keep the whole frame, centered, over a blurred background rather than cut someone off.

In Nysos this is automatic and decided shot by shot. The split screen kicks in when two people are clearly visible in a conversation shot: two 1080×960 panels, one face per panel. On a wide shot with three or four people, neighbors are grouped into the two panels. You can also reframe a single clip by hand from its page (one full-screen frame or two stacked frames) and switch back to automatic framing in one click.

If you're the one filming, one small thing helps people and tools alike: keep faces clear (no microphone in front of the mouth) and lit well enough to be tracked from frame to frame.

Why caption podcast clips, and where do the captions go?

Plenty of people watch short videos with the sound off: on a commute, at work, late at night. Without captions they scroll past in the first second. For a podcast it matters even more, because all of the content is in the speech.

On a split screen, the natural spot is the middle of the frame, right on the seam between the two panels: the eye travels from one face to the other and crosses the text on the way. That's where Nysos puts them in this case. Outside a split screen, position depends on the style: bottom for Nysos Pro, Storytime, Anime and Clean, center for Punch and Casual.

Captions are animated word by word, with the spoken word highlighted, karaoke style. There are six styles, or none at all; each style comes with its own colors, font and position. The text is corrected by AI after transcription; read over proper names and technical terms, and fix them on the clip's page. To keep one style across episodes, set it in a preset: if you don't choose, Nysos picks one for you.

What title keeps viewers past the first few seconds?

A title does two jobs: it tells people why to stop, and it doubles as the post title. It should give the stakes of the passage without giving away the payoff. “Why he quit EVERYTHING overnight” opens a question; “He quit his job” closes it.

For each clip, Nysos suggests a title of 60 characters or fewer, with one word in capitals, in the language of the video. It can be burned into the start of the clip as a title card: a white pill, black capitals, shown for 5 seconds, right where viewers decide whether to stay or swipe. You edit it on the clip's page before regenerating, so read it twice: it's the first line your audience reads. In a preset, the “Postable title” option turns it on for every clip. A title doesn't replace a strong first spoken line: if the clip opens on the question or the bold claim, the title only has to back it up.

How often should you post clips from an episode?

The right cadence is the one you can sustain for months. Three clips a week for three months beats twenty clips in release week and then silence. Nobody can promise you an audience result, and we distrust any numbered rule that claims otherwise.

One approach: for each episode, keep the handful of best passages and spread them across the following week, so the episode stays visible longer. Once your TikTok and YouTube accounts are connected, Nysos posts right away or, depending on your plan, at a date you pick, with a Planning view to move or cancel. The title can run up to 150 characters on TikTok and 100 on YouTube.

By hand or with Nysos: what actually changes?

By hand, the route is well known: relisten to the episode while spotting moments, cut in an editor, reframe, caption, export. You control everything, but the time adds up to tens of minutes per clip, a good share of it spent just finding the passages in an hour-long episode. That's workable for one or two clips a week.

With Nysos, you paste the episode link (YouTube, Twitch or Kick) or import the video file. The engine transcribes the speech, detects the language, finds the strong moments using the transcript, the audio and the picture, and offers them as vertical captioned clips. In practice, for one episode:

1. Paste the link or import the file, and pick a preset (caption style, title, intro).
2. Let processing run, then browse the suggested clips.
3. Keep only the ones that match one of the four shapes above.
4. Read each clip's start, end, title and caption text, and fix what needs fixing.
5. Download the MP4s (1080×1920, per clip or as a ZIP) or post straight to TikTok and YouTube.

Step 4 matters: Nysos proposes, you decide. Check where each clip starts and ends, read over the proper names, and ask whether the passage still makes sense out of context. Always review before you publish.

On cost, the free plan gives 60 source minutes a month with no credit card, rendered in 720p with a “Made with Nysos” watermark. Paid plans remove the watermark, and 1080p starts at Starter.

What source should you give Nysos for a podcast?

Nysos cuts video: the images are what let it frame faces. If your episode also exists as a video on YouTube, paste that link; otherwise, import the video file of your session. Filming your session, even with a basic camera, is what opens the door to clips.

The engine proposes the moments and you do the sorting: keep the clips you like, adjust the start and end, then publish.

Where should you start?

Take your latest episode, find three passages that fit one of the four shapes, and publish the best one. If you want to see the automated version, processing the same episode on the free plan gives you an honest point of comparison.