First: do you have video?
Everything about clipping a podcast follows from this one question, and it is worth answering honestly before reading any further advice.
If you filmed the episode, you are cutting video like anyone else – the only unusual part is that your shot almost certainly contains two people side by side, which vertical formats hate. If you recorded audio only, you cannot cut a video clip at all; you have to build one, and the thing you build is an audiogram.
Both routes are real and both work. What they share is the part that actually decides whether a clip lands: a podcast clip is carried by what is said. The picture is packaging. That is why a plain audiogram of a brilliant answer beats a beautifully shot clip of a dull one, and it is why the captions matter more here than on any other kind of clip.
When there is no video
An audiogram is a video made out of things that are not video: a still image, a waveform that moves in time with the audio, and captions. It exists because feeds will not carry a black rectangle, and it works because the waveform proves the thing is playing while the captions carry the meaning.
Keep it plain. The temptation is to add motion graphics and transitions, and they consistently make audiograms worse – the eye goes to the animation instead of the text. Cover art, a waveform, big legible captions, the show name and episode number, and nothing else.
The one thing worth spending effort on is the caption timing. An audiogram whose text lags the speech by half a second feels broken in a way viewers notice without being able to say why. Check the sync on the first ten seconds before you render the rest.
You can mark the in and out points directly on an audio file – browsers play MP3 and M4A, so the clipper will load your episode and let you fence off the moment before you build anything around it. Get the cut right first; the packaging is the easy part.
Two speakers, one vertical frame
If you filmed the episode, this is your only genuinely hard problem. A podcast shot is wide by design: two people, facing each other, with space between them. A vertical frame is the opposite shape, and something has to give.
Stacked is the safest default for a conversation. Both faces stay visible, nobody is cut off mid-sentence, and the viewer can follow who is reacting to what. It costs you image size, which matters less than it sounds because faces at half height are still legible on a phone.
Cutting to the active speaker gives the biggest, clearest picture and is the best option when one person is doing nearly all the talking – a monologue answer, a story, an explanation. It is also the most work, because every switch has to be right or the clip looks like it is arguing with itself. If you have separate camera angles per speaker, this becomes easy.
The wide shot letterboxed fits everything in and makes both faces too small to read. It always works and never works well. Use it when a clip is going out horizontally anyway – to X or LinkedIn – and choose one of the other two for vertical feeds.
Where the clips hide
Podcast episodes have a shape, and the same slots reward clipping in almost every show. Here is an hour laid out, with the verdicts you would expect on each.
- 00:00The cold openClip itIf you already pull a teaser for the top of the episode, you have done the work – that is a clip. Check whether it stands alone without the intro music behind it.
- 04:20The catch-up chatSkipWarm, human, and completely dependent on knowing both people. Nothing here survives being lifted out.
- 12:35The origin storyTighterUsually strong and usually slow. Cut into the middle of it – start at the turn, not at “so I grew up in…”.
- 23:10The question answered cleanlyClip itThe single most reliable podcast clip. One question, one answer, no cross-talk, done inside a minute. Cut the question if it is long and let the caption carry it.
- 31:04The disagreementClip itTwo people not agreeing holds attention far longer than one person being right. Start on the first objection, end before they reconcile.
- 38:50The long technical explanationSkipValuable and unclippable. Four minutes that depend on everything before them. Link a timestamp instead of butchering it.
- 46:12The recommendationClip it“The book that changed how I think about X.” Self-contained, specific, and endlessly shareable by people who have not heard the episode.
- 55:30The sign-offSkipFor the people already listening. Nothing in it for a stranger.
Notice that the skips are not bad content – the catch-up chat is often the most enjoyable part of an episode for regular listeners, and the technical explanation may be the most valuable. They are simply the parts that depend on everything around them, and a clip has none of that.
The single most reliable slot is the cleanly answered question, which is why interview shows clip more easily than co-host banter shows. If your format is two friends talking, expect fewer clips per episode and look harder for the disagreements and the recommendations. Choosing the best moments covers the judgment in general, and clipping content covers which formats reward the effort.
Cutting it
Once you have the moment, the mechanics are the same as any other clip and take about a minute: load the source, mark the in and out points, step frame by frame at both ends, preview, export. How to clip a video walks through it.
Two habits matter more for podcasts than elsewhere. Cut the run-up. Talk is full of throat-clearing – “yeah, so, I think…” – and almost every podcast clip improves by starting two seconds later than feels natural. End on the point, not the agreement. The other person saying “mm, totally” after a good line is where your clip should already have finished.
If you are producing several clips from one episode, load the source once and cut them all in a single sitting rather than returning to it five times. Turning long videos into short clips covers the batch workflow end to end.
Works on audio too
Load an episode – video or MP3 – and mark the moment in your browser
Captions are the picture
On a visual clip, captions are an accessibility feature and a concession to people watching on mute. On a podcast clip they are the content. Someone scrolling past sees two faces or a waveform – neither tells them anything. The text is the only thing that can make them stop.
Which means the usual advice to add captions understates it. Set them large, high-contrast and well clear of the interface, and read them before publishing: automatic captions mangle exactly the words a podcast clip depends on, which are names, jargon, book titles and numbers. A clip whose best line is transcribed wrong is worse than no clip.
Captions for video clips covers the tools and the placement. On vertical feeds, keep them out of the bottom band – the Reels guide has an interactive view of where the interface sits.
Send the guest their clips
This is the highest-return habit in podcast clipping and almost nobody does it properly.
When you publish an interview, cut two or three clips of the guest saying something they are proud of and send them the files directly – not a link to your post asking them to share it, the actual files, ready to upload. A guest who receives finished clips will usually post at least one, because you have removed all the work. A guest asked to reshare your post usually does not.
The reach maths is what makes it worth the extra twenty minutes: your clips reach people who already follow you, while the guest’s clips reach an audience that has never heard of your show. That is the only part of a podcast launch that compounds.
Two practical notes. Clip them, not you – the point is the guest looking good, so their best answer beats your best question. And send a clean file with no watermark and no burned-in show branding beyond a small credit, because heavy branding is the main reason guests quietly do not post the clip you sent.
If you do this for clients rather than your own show, it is a sellable deliverable in its own right – video clipping services covers packaging it.
A per-episode workflow
Roughly an hour per episode once it is a habit:
Scan the episode
Play at 1.5× and note timecodes for anything that registers. Do not judge yet – just collect.
Shortlist cold
Go back to the list and apply one test: does it stand alone with no context? Expect to lose half.
Cut everything in one sitting
Load the source once and export every keeper before closing the tab.
Package and send
Reframe, caption, schedule across the week – and send the guest theirs the same day the episode drops.
The scan is the expensive part, and it is nearly free if you edited the episode yourself: you already know where the good moments are, so the job collapses to writing down timecodes you remember. If you are clipping a back catalogue you have never listened to, that is the case where AI tools genuinely help – they produce a shortlist to react to, even if you re-cut the ends yourself.
FAQ
Can I make clips if my podcast is audio only?
Yes – that is what an audiogram is for: cover art, a moving waveform and captions over your audio. It will not compete with a well-shot video clip on looks, but the words are what carry a podcast clip anyway. You can mark the in and out points straight on an audio file in the clipper, since browsers play MP3 and M4A.
How many clips should one episode produce?
Three to five good ones from an hour is realistic. More than that usually means the bar has slipped. Publishing four clips people watch beats publishing eight they scroll past, because the weak ones teach your audience to ignore your name.
Should I clip the question or just the answer?
Start with just the answer and see if it stands. If it does not make sense without the question, either include a trimmed version of the question or put it in the caption. A long question read aloud at the start of a clip is the most common reason podcast clips lose people.
Do I need a separate vertical camera setup?
No. A wide two-person shot can be reframed, and this page covers the three ways to do it. A vertical-first setup makes clipping easier, but it is an optimisation, not a prerequisite – and it makes the long-form episode look worse. Horizontal to vertical video covers the reframing.
What length works for podcast clips?
Usually 30 to 60 seconds – longer than a typical visual clip, because the content is an argument rather than a moment. If the point takes 90 seconds to land properly, let it, but check that the first ten seconds earn the next eighty.
Is it worth clipping guest episodes specifically?
Yes, and it is the most underused growth route in podcasting. A guest who receives three ready-to-post clips will usually post at least one, which puts your show in front of an audience that has never heard of it. See below for how to package them.
Related: turning long videos into short clips, YouTube videos to Shorts and earning from clipping.