Every podcaster has heard "you should be clipping your episodes." Almost nobody does it well. Doing it well is the difference between a podcast with 500 listens per episode and one with 500,000. Here's the exact workflow the top-performing indie podcasts (My First Million, Lenny's Podcast, Modern Wisdom) use in 2026 — reverse-engineered from our teardown of 40 episodes.
The one rule: clips are not summaries
The single biggest mistake: trying to make each clip "about" what the episode is about. Clips work when they stand alone as complete, quotable moments — the topic is irrelevant.
A great clip has one of these three shapes:
- A specific number that surprises ("I spent $47,000 on this and made $12").
- A contrarian take ("Actually, cold outreach still outperforms warm intros in 2026").
- A short story with a payoff ("The exact moment I knew we had product-market fit").
If a clip needs setup to make sense, cut it.
The extraction workflow
Step 1: AI-transcribe with speaker labels
Use Whisper (open-source) or Descript. Get a timestamped transcript with speaker diarization.
Step 2: Feed the transcript to a clip-scoring model
Prompt template:
> "You are a viral clip finder. From the transcript below, identify every moment that meets one of these three tests: (a) contains a specific surprising number, (b) contains a contrarian take that violates conventional wisdom, (c) contains a self-contained mini-story with a clear payoff. For each moment, return: start_time, end_time (30-90 seconds), a one-sentence hook, and a virality score 1-10. Return only moments scoring 7+."
This yields 15–40 candidates per 60-minute episode.
Step 3: Human filter to top 12
You cannot ship 40 clips per episode without diluting reach. Pick the 12 with the highest scores AND the most distinct topics.
Step 4: Cut with the "3-second rule"
Every clip must be re-cut so the interesting moment happens in the first 3 seconds. Trim any "so, uh, one thing I've been thinking about…" preamble. Start with the punchline sentence.
Step 5: Captions and B-roll
Auto-captions using Opus Clip, Captions.ai, or Submagic. Add:
- Karaoke-style word highlighting.
- One B-roll insert per clip (relevant meme, chart, or footage).
- A static "hook headline" text overlay for the first 3 seconds.
Step 6: Platform-native repurposing
Do not cross-post the same file. Each platform gets a tuned version:
- TikTok/Reels: 9:16, 40–60 seconds, hook overlay top-third (avoid caption box), heavy captions.
- YouTube Shorts: 9:16, 45–75 seconds, hook slightly higher, less overlay clutter.
- X: 9:16 or 1:1, keep under 45 seconds (X's completion cliff), open with the punchline visible in the still-frame preview.
- LinkedIn: 1:1, 30–60 seconds, softer opening line (LinkedIn hates over-produced hooks).
Cadence
The top podcasters ship 3–5 clips per platform per week (not per episode). One 60-minute episode fuels 3 weeks of shorts.
Tooling stack
- Cutting: Opus Clip Pro or Descript.
- Captions: Submagic (best value) or Captions.ai (best AI features).
- Scheduling: Buffer, Metricool, or Later.
- Analytics: Native platform + one aggregator (Metricool works well).
Full end-to-end: 90 minutes of human time per 60-minute episode. That's the actual leverage point that makes podcast growth possible in 2026.