Most podcast growth advice ends the same way: "post clips on Instagram and TikTok." Useful, until you sit down to do it and realize a 60-minute episode gives you no obvious place to start. Which 40 seconds? Why that one? And how do you make a horizontal recording look right in a vertical frame without lopping off whoever's talking?
What follows is a practitioner's workflow for turning a podcast into reels, the kind you can run every week without it eating your Sunday. It's for hosts who already record episodes and want short clips that earn watch time, not clips that just fill a content calendar. There's no single viral hack here. Just the repeatable steps that separate clips people finish from clips people scroll past.
What makes a podcast clip work (and what doesn't)
A reel is not a shorter episode. It's a self-contained moment that makes sense to someone who has never heard your show. That one distinction drives every decision below.
The clips that perform tend to share three traits:
- A clear hook in the first two seconds. Viewers decide almost instantly whether to keep watching. A clip that opens with "so, um, anyway" has already lost half of them.
- One idea, fully contained. If a viewer needs the previous five minutes to follow along, it won't land. The best moments stand on their own: a story, a strong opinion, a fact that makes you go huh.
- A reason to react. People share clips that make them feel something. Surprised, validated, amused, or annoyed enough to argue in the comments.
What doesn't work? Clipping something because it was "a nice bit," posting the raw horizontal video with black bars top and bottom, or leaving captions off. Instagram and TikTok get watched on mute far more often than people admit. Text on screen is what stops the thumb.
Step 1: Find the moments while you still remember them
The hardest part of clipping isn't editing. It's finding the 30-to-60-second windows worth editing at all. Scrubbing an hour of audio after the fact is where most people quietly give up.
Two ways to make it painless:
- Mark moments live. The second a guest says something that makes you think "that's the clip," jot the timestamp. Sticky note, Notes app, back of your hand, doesn't matter. Ten seconds of effort during the recording saves you an hour of scrubbing later.
- Work from the transcript, not the timeline. Reading is far faster than listening. Skim the transcript for complete thoughts, strong claims, and story beats, then pull the timestamps from there.
Aim for three to five candidate moments per episode. You won't publish all of them, and you shouldn't try. A small pool of strong options beats one clip you forced because you felt you needed something. Look for the counterintuitive take, the concrete number, the short personal story, the flash of genuine disagreement, or a crisp one-line answer to a question your audience actually asks.
Step 2: Cut tight, then cut tighter
Once you have a moment, trim it to only what the clip needs. A good podcast reel usually runs 20 to 60 seconds. If the setup eats fifteen seconds before the payoff lands, either cut into the setup or pick a different moment.
A structure that holds up:
- Hook (0-2s): open on the strongest line, not the run-up to it. Starting mid-sentence is completely fine if that sentence is the hook.
- Payoff (middle): the story, the point, the reveal.
- Button (last few seconds): end on a clean line, not a trailing "yeah, so, anyway." Cut the instant the thought completes.
Don't be precious about chronology. You can open with the punchline and let the rest of the clip explain how you got there. The timeline of the recording is not the timeline of the clip, and pretending otherwise is why so many clips sag in the middle.
Step 3: Reframe for vertical without decapitating anyone
Most guides wave past this step. It's also the one that makes clips look amateur when it's done wrong. Your episode was almost certainly recorded in horizontal 16:9, and reels are vertical 9:16. That's not a crop you set once and forget. The person talking moves around, and two people almost never sit dead center in the frame.
What good vertical framing actually requires:
- Follow the active speaker. When the host talks, the frame belongs on the host. When the guest answers, it moves to the guest. A static center crop leaves you staring at the empty gap between two heads.
- Keep faces in the upper-middle third. On a phone, the bottom of the screen is buried under captions, the account handle, and the like and comment buttons. Frame heads high enough that the interface doesn't slice through them.
- Don't over-cut. Jumping between speakers on every syllable is dizzying. Switch on turns in the conversation, not on every back-and-forth "mhm."
Doing this by hand is genuinely tedious. Keyframing the crop position for a two-person conversation is the part of clip-making people quietly dread, and I don't blame them. It's also exactly the kind of repetitive, rules-based work automation handles well, which we'll come back to.
Step 4: Caption everything
Captions are not optional. A large share of short-form video gets watched with the sound off, and on-screen text is both an accessibility baseline and the thing that holds a distracted viewer in place. Auto-generated captions get you most of the way there. Always proofread them anyway. Names, jargon, and numbers are precisely what automatic transcription fumbles, and a wrong figure in a clip that's supposed to be about results quietly torpedoes the whole point.
A few things that matter:
- Keep captions large and high-contrast, sitting in the safe zone above the interface buttons.
- Show only a few words at a time so the eye can keep pace.
- Fix proper nouns and figures by hand. The rest can be lightly cleaned and left alone.
Step 5: Write a hook caption and post on a schedule
The text caption you write in the post matters nearly as much as the video. Lead with a line that frames the clip. A question, a claim, or the tension in the moment beats "Great chat with @guest on this week's episode" every time. Add a short call to action ("full episode in bio") and a handful of relevant tags. Then post consistently. Three solid clips a week, every week, beats a burst of twelve followed by three weeks of silence, because the algorithm and your audience both reward showing up.
Where AI fits in this workflow
Everything above is doable by hand. The reason most podcasters don't keep it up is that the work is repetitive: scrub, trim, keyframe the crop, caption, proofread, then do it four more times per episode. That's precisely the kind of grind AI is good at removing.
Liyam AI is built around the podcast conversation itself, and it touches several of these steps directly. During recording, it flags standout moments as they happen, so you finish an episode with your clip candidates already marked instead of an hour of scrubbing ahead of you. Afterward it produces an accurate transcript you can skim for moments, and it handles the vertical reframing, auto-cropping 16:9 footage to 9:16 and following the active speaker so the frame stays on whoever is actually talking. The point isn't to replace your judgment about which moments matter. It's to strip out the tedious middle so you can get from "that was a good line" to a captioned, correctly framed reel without dreading the whole process.
Pick the step that costs you the most time and let automation take that one first. For most hosts that's the reframing and captioning grind. Automate it, and clipping stops being the chore that quietly falls off your weekly list.
FAQ
How long should a podcast reel be?
Most land between 20 and 60 seconds. Shorter clips, under 30 seconds, tend to get higher completion rates, which the algorithms reward. Let the moment decide. Cut the instant the thought is complete instead of padding to hit a number.
Do I need a video podcast to make reels?
It helps, but no. With video, you can crop and reframe it into vertical clips. Audio only? You can still make reels as audiograms, a waveform or static image with captions over the audio, and they perform well when the moment and the text hook are strong.
How many clips should I make per episode?
Pull three to five candidate moments per episode and publish the two or three strongest. Consistency beats volume. A steady two or three clips a week does more for you than a big batch followed by a gap.
What's the best way to find clip-worthy moments?
Mark timestamps live while recording, every time something makes you think "that's the clip," then skim the transcript afterward for complete thoughts and strong lines. Reading the transcript is much faster than re-listening to the whole episode.
The takeaway
Turning a podcast into reels isn't a creative mystery. It's a repeatable process: find complete moments, cut them tight, frame them properly for vertical, caption them, then post on a schedule you can actually keep. The hosts who grow on short-form aren't the ones with a viral trick. They're the ones who made the process light enough to run every single week.
If the clipping grind is what keeps you from being consistent, see how Liyam AI handles the moment-flagging, transcription, and vertical reframing, so the human part, the conversation, stays yours.