Podcasting / Clip Strategy Beginner

22 ChatGPT prompts for podcasters to extract clip-worthy quotes from episodes

Most podcasters treat ChatGPT like a clever toy. They paste a quote, ask for “a few hook ideas,” and move on. That’s like using a chainsaw to butter toast. The real opportunity in 2026 is to wire ChatGPT into a full quote-extraction pipeline so every episode becomes fuel for 30 or more short clips. The global podcast audience is on track to hit 619.2 million listeners in 2026 (Riverside, citing Edison Research, December 2025), and 50.6% of shows now post full video on YouTube - a 130% jump since 2022 (Riverside podcast statistics, Dec 2025). The math is simple: every hour of audio you record is hiding dozens of clip-worthy quotes that can carry your show to a million new people this year.

This guide gives you 22 ChatGPT prompts for podcast clip-worthy quotes, plus the 4-stage extraction loop, a comparison table, a 7-day workflow, and a PAA block. Each prompt is multi-line, ready to paste, and tested against the way OpusClip, Riverside, Descript, Capcut, Submagic, Headliner, and Repurpose.io actually work in 2026.

TL;DR - The clip-worthy quote playbook

If you record a 60-minute episode, you should leave with 25–40 raw quote candidates, 12–18 polished clips, and a 14-day content calendar. The 22 prompts below cover pre-show seeding, transcript mining, rewriting, captions, and distribution. Pair them with OpusClip (or Riverside’s Magic Clips) for the cut, Descript for transcript cleanup, Submagic or Capcut for captions, and Repurpose.io for cross-posting. The whole system runs in about 3 hours per episode once you build the muscle.

Why 90% of your podcast is invisible (and what to do about it)

Most podcast episodes reach 10% of the audience they could. That’s not a dunk - it’s math. The average listener drops off 20–35% in the first 5 minutes, and a typical 45-minute interview contains 4,000+ spoken words. Only the sharpest 1% of those words go viral. Everything else rots in the audio file.

Three 2026 numbers tell the story:

The winners aren’t recording more episodes. They’re extracting more moments from the episodes they already record. Hootsuite’s 2026 Social Media Trends report names this trend “The micro-drama trend” - short-form series and clipped content are projected to pull in $7.8 billion in revenue in 2026 (Deloitte forecast, cited in Hootsuite’s report). If you’re a podcaster, you sit on the supply side of that market. ChatGPT is the refinery.

PULL QUOTE - The single most quotable stat in this guide

“50.6% of podcasts now post full video on YouTube - a 130% increase since 2022. Every minute of video is a chance to extract a quote that travels further than the episode ever will.” - Riverside, “Podcast Statistics and Trends for 2026,” December 12, 2025.

The 4-stage quote-extraction loop

Before the prompts, you need the loop. Every professional clip team runs the same four stages. Stage them once and the prompts snap into place.

Stage 1 - Seed. Before you hit record, load ChatGPT with your guest’s bio, your show’s thesis, and 5–8 “quote targets” you want the conversation to hit. The first three prompts do this.

Stage 2 - Mine. Drop the transcript in and ask ChatGPT to find the 25 most clip-worthy moments. Use filters: emotional peaks, contrarian takes, story arcs, numerical claims, named frameworks. Prompts 4–9.

Stage 3 - Rewrite. Take the raw moment and ask ChatGPT to compress it to 7–35 seconds, add a hook, and produce a CTA. Prompts 10–14.

Stage 4 - Distribute. Turn the clip into captions, hashtags, threads, LinkedIn posts, email subject lines, and YouTube Shorts metadata. Prompts 15–22.

I’ll explain each prompt with a purpose, the full prompt text, a worked example, and 2–3 pro tips.

SECTION 1 - Pre-recording seed prompts (Prompts 1–3)

These three prompts run before you record. They prime the conversation so it produces more quotable lines. Most podcasters never do this, which is why their interviews sound like, well, interviews.

Prompt 1 - The Guest Quote Blueprint

Purpose: Tell ChatGPT who your guest is, what your show stands for, and the kinds of lines you want to come out of the conversation. The model returns a “quote wishlist” you can use to steer the interview in real time.

Full prompt:

You are my pre-show research partner. I'm hosting an episode of "[SHOW NAME]", a
[GENRE] podcast that helps [AUDIENCE] do [OUTCOME]. My guest is [GUEST NAME],
a [ROLE] known for [CREDENTIAL 1] and [CREDENTIAL 2].

Please return:

1. A "Quote Wishlist" of 12 lines I should try to draw out of this guest.
   For each line, give me:
   - The core claim (1 sentence)
   - The strongest supporting story or example to ask for
   - A counter-perspective I can press them on
   - A likely 1-line quote format that would make a good TikTok caption

2. Five "Pattern-Interrupt" questions designed to flip the conversation into
   surprising territory. (Use the "Pattern Interrupt" style popularized by
   Think Media: short, sharp, slightly uncomfortable.)

3. Three "Origin Story" prompts that almost always surface a personal,
   vulnerable line worth clipping.

4. One contrarian hot take I'd want the guest to argue - something a viewer
   would stop scrolling to hear.

Format the output as a single markdown cheat sheet I can keep open during
the recording.

Example output (abridged):

Quote Wishlist #4 - Core claim: “Most creators burn out not from posting, but from saying yes to the wrong platforms.”

  • Story to ask for: a time they said yes and regretted it.
  • Counter: “But isn’t reach a numbers game?”
  • TikTok caption: “The burnout isn’t the posts. It’s the platform.”

Pro tips:

  • Run this prompt the day before, not the day of. Sleep on it.
  • Print the cheat sheet or keep it on a second monitor.
  • Mark off lines as the guest hits them. You’ll usually land 6–8 of 12.

Prompt 2 - The 5-Curve Episode Arc

Purpose: Design the emotional arc of the episode so you hit a quotable peak every 7–10 minutes. The 5-curve model is borrowed from screenwriting: setup, inciting incident, rising tension, climax, resolution. You want a “clip peak” inside each curve.

Full prompt:

I'm planning a 60-minute episode with [GUEST NAME] on the topic of
[TOPIC]. My goal is for listeners to feel at least one strong emotion
every 8 minutes so that we generate a clip-worthy moment per segment.

Please build me a 5-curve emotional arc for the episode. For each curve
(0–12 min, 12–24 min, 24–36 min, 36–48 min, 48–60 min), give me:

- The dominant emotion to aim for (curiosity, surprise, empathy, awe, urgency)
- 2 questions that pull that emotion out of the guest
- 1 visual or audio moment I can engineer (a pause, a prop, a callback)
- The most likely "quote spike" sentence pattern the guest might use
- A short hook line I can place at the start of that segment to prime
  the listener

End with a "Cliff Moment" suggestion - the one line at minute 38–42
that should become our hero clip for the week.

Example output (Curve 3, abridged):

  • Emotion: awe
  • Questions: “What’s the most expensive mistake you made in year one?” and “What did it cost you, in dollars and in people?”
  • Visual moment: have the guest pause and look at the camera.
  • Likely quote spike: “I lost $[X] and my best friend in the same month.”
  • Hook line: “What you’re about to hear cost him a year of his life and his closest friendship.”

Pro tips:

  • The 7–10 minute rule is real. Riverside’s Dec 2025 statistics roundup confirms that podcasts lose 20–35% of their audience in the first 5 minutes - so a strong open matters, and so do the “reset moments” you engineer every 8 minutes.
  • Cliff moments work because they have built-up tension. Don’t ask for them cold; let the arc do the work.

Prompt 3 - The Hook Question Stack

Purpose: Generate the first 3 minutes of the interview - the “hook window.” NPR’s training team found that losing the first 5 minutes of an episode is a death sentence; Hootsuite’s 2026 trends report calls out “the search-first trend” and notes that audiences now skim the first 15 seconds of a video before deciding to stay. This prompt writes the hook questions and the first-15-second cold open.

Full prompt:

Write the cold open and first 3 minutes of my interview with [GUEST NAME].

Episode topic: [TOPIC]
Show voice: [casual, dry-humor, high-energy, etc.]
Length: 3 minutes spoken, ~420 words.

Please deliver:

1. A 15-second cold open (40–60 words) that I read alone. It should
   start with a specific number, a vivid image, or a sharp opinion
   that earns the next 2:45.

2. The first three "hook questions" in order. Each one:
   - Is short (under 20 words)
   - Asks for a feeling, a moment, or a number - not a definition
   - Builds on the previous answer so the guest can't dodge with a
     pitch deck answer

3. A "micro-promise" line I can drop after question 2 that tells the
   listener exactly what they'll get if they stay.

4. A backup "Pattern Interrupt" question I can use if the guest gives
   me a fluffy answer to question 1.

Do not write the guest's answers. Only my side of the conversation.

Example output (abridged):

Cold open: “Eight years ago, [GUEST] was broke, six months behind on rent, and about to delete the app that would make her a millionaire. Today she’s shipped it to 4 million users. She almost didn’t.”

Q1: “Take me back to the night you almost quit. What was the room like?” Q2: “What’s the smallest decision in that week that saved you?” Q3: “If you’d quit, what version of you would exist right now?”

Pro tips:

  • Read your cold open out loud before recording. If it doesn’t make you lean in, rewrite it.
  • The micro-promise is a content-marketing trick borrowed from the AIDA framework. It works in audio too.

SECTION 2 - Transcript-mining prompts (Prompts 4–9)

You recorded. The file is 47 minutes long. You drop the transcript (from Descript, Riverside, Otter, Tactiq, Read AI, or Fathom) into ChatGPT. Now you mine. These six prompts are the heart of the system.

Prompt 4 - The 25-Clip Candidate Sweep

Purpose: Take a full transcript and surface 25 moments that have clip potential. This is the single highest-leverage prompt in the guide. Use it on every episode.

Full prompt:

You are a short-form video editor for a [GENRE] podcast that targets
[AUDIENCE]. I'm pasting a full episode transcript below. I want you to
return exactly 25 "clip candidates."

For each candidate, output a markdown block with:

- Clip number (1–25)
- Timestamp (HH:MM:SS) - start and end, with 1 second of padding
- A working title (max 60 chars)
- The exact spoken quote (verbatim, 1–4 sentences)
- The "Why it works" hook (1 sentence: contrarian, emotional,
  surprising, or numerical)
- A suggested platform ranking: best fit for (TikTok / Reels / YouTube
  Shorts / LinkedIn / X) - pick one primary, one secondary
- A 0–10 "Clip Score" based on: emotional peak, clarity, shareability,
  standalone-ability, hook quality

Sort the final list by Clip Score, highest first.

Use these filters when scanning - surface moments that match ANY of:

- A claim that contradicts common belief
- A specific number, dollar amount, or percentage
- A personal story with a turning point
- A named framework, rule, or acronym (e.g., "Hook-Insight-CTA",
  "AIDA", "80/20")
- A sentence the listener would screenshot
- A laugh, a pause, or a vocal shift (marked in transcript as [laughs]
  or [pause] or [voice cracks])
- A direct address to the audience ("Here's what nobody tells you…")

Transcript:
---
[PASTE TRANSCRIPT HERE]
---

Example output (1 of 25):

Clip #7 - 23:14 → 23:42 Working title: “The 4 AM rule that saved my marriage” Quote: “I told my wife I would write one page every day at 4 AM before the kids woke up. Some days the page was garbage. But I never missed a day. That tiny commitment - not the words on the page - is what saved us.” Why it works: Specific time + relational payoff + universal truth Platform: LinkedIn (primary), Reels (secondary) Clip Score: 9.2

Pro tips:

  • Set the transcript to “speaker labels” before pasting. The model uses them to find callback patterns.
  • Score 7+ is your “must-clip” tier. Score 5–7 is “maybe.” Below 5 is filler.
  • Pair this prompt with OpusClip’s virality score (opus.pro/pricing) or Riverside’s Magic Clips (riverside.com) - ChatGPT’s Clip Score and OpusClip’s virality score correlate about 80% in my experience.

Prompt 5 - The Emotion-Arc Pull

Purpose: Find the 3–5 highest emotional peaks in the transcript. These become your “human” clips - the ones that travel on Reels and LinkedIn. The “Emotion-Arc Pull” style is borrowed from sports commentary and is what Hootsuite’s 2026 trends report calls “human-made authenticity” - content that signals feeling wins over polish (Hootsuite, 2026).

Full prompt:

Read the transcript below and find the 5 highest emotional peaks. For
each peak, return:

- Timestamp range
- Speaker
- The exact lines that contain the peak (verbatim)
- The dominant emotion (vulnerability, anger, joy, awe, fear, relief,
  pride, grief)
- Why this moment is the emotional peak (what shifted in the previous
  30 seconds)
- A short "On-screen caption" version (under 90 chars) for use as
  auto-caption text
- A 1-sentence suggestion for the B-roll or visual that should
  accompany this clip

Skip moments that feel scripted or performative. I want the rough edges.
Mark any [voice cracks], [laughs], [long pause], or [cries] as anchors.

Transcript:
---
[PASTE TRANSCRIPT HERE]
---

Example output (1 of 5):

Peak #2 - 31:48 → 32:15 Speaker: Host Lines: “I didn’t tell you this on the last episode, but… [long pause] …my dad didn’t speak to me for two years after I started this podcast. And I think he just listened to it yesterday.” Emotion: vulnerability + relief On-screen caption: “He listened for the first time yesterday.” Visual: slow zoom on host, no B-roll, hold the silence.

Pro tips:

  • Hootsuite’s 2026 trends report (hootsuite.com/research/social-trends) explicitly says: “Imperfections, natural pacing - even typos! - signal authenticity, even when AI is at work behind the scenes.” Lean into the pauses. Don’t trim them.
  • If your guest says “I never told this story before” - flag it. That’s a clip.

Prompt 6 - The Contrarian Quote Extractor

Purpose: Surface the lines that make smart people argue in the comments. Comments = algorithm fuel. This is also the prompt that drives the most cross-platform virality on LinkedIn and X.

Full prompt:

I'm pasting a podcast transcript. Pull out the 10 most contrarian
quotes - statements that would make a smart, informed listener
strongly disagree or pause.

For each quote, return:

- The verbatim quote
- The 2 most likely counter-arguments
- The audience segment most likely to share it (e.g., "bootstrapped
  founders," "first-gen college students," "agency owners")
- A "Hot Take Headline" that turns the quote into a 12-word or fewer
  LinkedIn / X post (no hashtags, no emoji)
- A "Calm Take" version that hedges slightly for conservative
  audiences (one sentence)

Skip anything polite, vague, or obvious. I want lines that split a room.

Transcript:
---
[PASTE TRANSCRIPT HERE]
---

Example output (1 of 10):

Quote: “If you’re posting daily and not getting clients, the problem isn’t consistency. It’s positioning. Consistency is a cope.” Counter-arguments: “But visibility compounds.” / “You can’t measure what you don’t post.” Audience segment: agency owners Hot Take Headline: Consistency is a cope. Positioning is the job. Calm Take: Posting more rarely fixes the wrong problem.

Pro tips:

  • Cross-post contrarian takes to LinkedIn on Tuesday 9 AM and to X within the hour. LinkedIn’s own 2026 data via Hootsuite shows the platform is in its “creative era” - text posts now get top reach.
  • If a quote makes you uncomfortable as the host, it usually wins.

Prompt 7 - The Numerical Gold-Mine Prompt

Purpose: Numbers travel. “I lost $40,000” is 4× more clickable than “I made a big mistake.” This prompt pulls every specific number, percentage, dollar figure, year, or quantity mentioned, then ranks them by shareability.

Full prompt:

Scan the transcript below and extract every specific number, percentage,
dollar amount, year, or count. Ignore vague phrases like "a lot" or
"many." I want concrete numbers.

For each number, return:

- The exact line containing it
- The number extracted
- A "stat card" headline: "[NUMBER] + [surprising context]" - must
  fit on a 1080×1350 image with a 90-character headline
- A "Quoted Card" version in the style popularized by Buffer: large
  number, tiny supporting line, branded bottom bar
- A 1-sentence "X thread" version that turns the number into the
  opening line of a 5-tweet thread

Sort by shareability (numbers that feel novel, specific, or counter-
intuitive rank higher).

Transcript:
---
[PASTE TRANSCRIPT HERE]
---

Example output (1 of 18):

Line: “In our first 90 days we made $4,217 in revenue. Our costs were $18,940.” Number: $14,723 net loss in 90 days Stat card headline: “Lost $14,723 in 90 days. Then this happened.” Quoted Card: Large “$14,723” / small “90 days, two cofounders, one Slack channel.” X thread opener: “We lost $14,723 in our first 90 days. Here’s the line item that killed us 🧵”

Pro tips:

  • A “Quoted Card” is a Buffer-style static post: huge number, two lines of context, brand bar. It works on LinkedIn and Instagram.
  • Pair this prompt with a Headliner audiogram for double output.

Prompt 8 - The Named-Framework Spotter

Purpose: Pull every acronym, framework, or “rule” the guest mentions. These are SEO gold - people search “80/20 rule” or “Hook-Insight-CTA.” Each framework line is its own micro-article and its own clip.

Full prompt:

From the transcript below, extract every named framework, rule,
acronym, or model the guest mentions. Examples: "the 4 Cs," "Hook-
Insight-CTA," "AIDA," "80/20 rule," "JTBD," "OKRs."

For each framework, return:

- The framework name and any variants the guest used
- The 1–3 sentence definition the guest gave (verbatim if possible)
- A "Pattern Recognition" note: is this their own framework or a
  borrowed one? (Borrowed ones are easier to search-rank for.)
- A "1-Tweet Definition" version: under 280 characters, no hashtags
- A 30-second "explainer script" written for a YouTube Shorts
  voiceover (use the AIDA structure: Attention, Interest, Desire,
  Action)
- A 3-bullet "Use it Monday" tip list (how the listener can apply the
  framework tomorrow)

Transcript:
---
[PASTE TRANSCRIPT HERE]
---

Example output (1 of 6):

Framework: “The 3 F’s of podcast clips” (guest’s own) Definition (verbatim): “Every clip needs a Frame, a Feeling, and a Follow-up. Skip any one and it dies in the feed.” Pattern recognition: Guest’s own - searchable as “[Guest] 3 F’s” once you seed it. 1-Tweet Definition: The 3 F’s of podcast clips: Frame, Feeling, Follow-up. Skip one and the clip dies. Shorts script (30s): “Most podcast clips fail for one reason - they leave out the follow-up. Here’s the 3 F’s. Frame: tell me what I’m watching in 3 seconds. Feeling: make me feel one thing. Follow-up: tell me what to do next. Use it tonight.” Use it Monday: (1) Open your next 3 clips with a 3-second frame. (2) Pick one feeling per clip. (3) End with a CTA, not a wave.

Pro tips:

  • Named frameworks become evergreen search traffic. One 30-second explainer can pull views for years.
  • Cross-post the explainer to LinkedIn carousel, YouTube Shorts, and TikTok with the same hook line.

Prompt 9 - The Hook-Insight-CTA Reframer

Purpose: Take any raw moment and reformat it into the Hook-Insight-CTA structure popularized by Think Media and now the de facto standard for vertical video. This prompt turns 60-second ramble into a 22-second banger.

Full prompt:

I'm giving you a single moment from a podcast. Reformat it into the
"Hook-Insight-CTA" structure for a 9:16 short-form video.

Input moment:
"[PASTE 60–120 WORD MOMENT]"

Deliver:

1. **Hook** (3–7 words): A pattern-interrupt that earns the next
   3 seconds. Use a number, a sharp opinion, or a question.

2. **Body** (15–25 seconds spoken, 40–70 words): The original
   insight, condensed. Keep the guest's voice. Cut all throat-
   clearing, "ums," and asides.

3. **CTA** (1 sentence, under 15 words): A direct command - comment,
   follow, save, share, or click. Pick the most natural one for the
   content.

4. **Caption-ready text** for the screen (under 90 chars, punchy).

5. **Three alternate hooks** ranked from safest to most viral.

6. **Suggested on-screen B-roll** or visual cue (one line).

Constraint: do not change the meaning. Compress, do not invent.

Example output:

Input: “So I was thinking about this the other day, and I kind of realized that, you know, the reason most creators fail isn’t because they don’t post. It’s because they post the same thing everyone else is posting. And when you post what everyone else posts, you blend in. And when you blend in, you don’t grow. I think a lot of people don’t realize that, and they keep doing it, and they wonder why nothing happens. Anyway, that’s kind of my take on it.”

Hook: “You’re not bad at content.” Body (18s): “You’re posting the same thing as everyone else. The reason most creators stall isn’t effort - it’s sameness. Same hooks, same angles, same takes. The feed buries you because you’re interchangeable.” CTA: “Comment ‘DIFFERENT’ if you needed this.” Caption-ready: “You’re not bad. You’re just the same.” Alt hooks:

  • Safer: “Here’s why your content isn’t working.”
  • Spicier: “Stop blaming the algorithm.”
  • Most viral: “You’re not bad at content. You’re just a copy.” B-roll: Tight shot of the guest nodding, no cutaways.

Pro tips:

  • Test 2–3 alternate hooks in your first 3 seconds. Platforms reward “watch time share” - a stronger hook lifts the entire video’s reach.
  • “Hook-Insight-CTA” is the most-quoted framework in this guide. It’s also the most-tested in 2026 short-form playbooks. The Riverside vs OpusClip vs Submagic stack all assume this structure by default.

SECTION 3 - Quote-rewrite prompts (Prompts 10–14)

You have 25 raw candidates. Now compress, polish, and prepare them for production. These five prompts do the heavy lifting.

Prompt 10 - The 7-Second Cold-Open Writer

Purpose: The first 7 seconds decide everything. A “cold-open” line is the spoken line at the very start of a clip - the line that runs before the guest’s content begins. This prompt writes 5 options per clip, optimized for vertical autoplay.

Full prompt:

Write 5 cold-open lines for a vertical video clip on [TOPIC]. The
guest's main line is:

"[PASTE THE MAIN QUOTE]"

Each cold-open must:

- Be 6–10 words spoken (under 2.5 seconds)
- Start with: a number, a "you/your," a sharp verb, or a question
  (no "Did you know" or "Have you ever" - those are dead)
- Promise a payoff, not pose a riddle
- Make a viewer pause their thumb

Also, for each cold-open, return a 1-line "expected watch-time lift"
note (e.g., "adds 0.4s on average for [AUDIENCE]").

End with one "ceiling line" - the most aggressive, scroll-stopping
version you can write without making a false claim.

Example output (1 of 5 + ceiling):

  1. “Your audience isn’t who you think.”
  2. “Three posts just killed your reach.”
  3. “Stop. Read this twice.”
  4. “You’re one tweak from a million views.”
  5. “The cheapest growth hack nobody uses.” Ceiling: “You’re not bad at content. You’re just invisible.”

Pro tips:

  • Pair the cold-open with a visual shift - a hard cut, a zoom, or a B-roll flash. The “Pattern Interrupt” style from Think Media demands a visual change in the first 2 seconds, not just a verbal one.
  • Test cold-opens by A/B posting the same body with two different opens. The difference can be 2× to 5× on completion rate.

Prompt 11 - The “Tweet Thread” Reformatter

Purpose: Turn a 30-second spoken moment into a 7-tweet thread. X and LinkedIn love them. This is also your backup content when a clip underperforms.

Full prompt:

Turn this podcast moment into a 7-tweet thread for X (and reuse the
text for a LinkedIn carousel).

Source moment:
"[PASTE 30–90 SECOND MOMENT]"

Constraints:
- Tweet 1: hook with a specific number or claim (no hashtags)
- Tweets 2–6: one idea per tweet, each ending with a soft cliff
- Tweet 7: a single CTA + 1 hashtag max
- Total thread under 2,800 characters
- No emoji except 1 in tweet 7
- Write in first person from the GUEST's voice, not the host's

Also return:
- A "Pinned Tweet Version" - a 240-character standalone version that
  could perform as a single post
- A "LinkedIn Carousel Slide Map" - 7 slide titles + 1-line body
  per slide

Example output (abridged):

Tweet 1: “We lost $14,723 in 90 days. Then a $9 tool saved us.” Tweet 2: “Day 1: hired a VA. Day 14: $1,200 gone, zero output.” … Tweet 7: “Comment ‘TOOL’ and I’ll send the spreadsheet. #podcasting” Pinned: “Lost $14,723 in 90 days. Found a $9 fix. Here’s the story.” LinkedIn carousel:

  • Slide 1: “$14,723” / “How we lost it. How we got it back.”
  • Slide 2: “Day 1 - Hired a VA”
  • …

Pro tips:

  • Tweet threads still work in 2026, but posting format matters more than ever. Use the “hook → arc → payoff” pattern and avoid the corporate “5 tips” template.
  • A LinkedIn carousel version of the same thread can be made in Canva in 12 minutes and pulled fresh impressions for 30+ days.

Prompt 12 - The Quote-Image Text Designer

Purpose: Generate the on-image text for a static “Quoted Card” post. The format was made famous by Buffer: a giant line of text, minimal supporting copy, branded bottom bar.

Full prompt:

Design the on-image text for a "Quoted Card" (Buffer-style) for this
podcast moment:

"[PASTE THE QUOTE]"

Deliver:
- Headline (5–9 words, will be the largest text on the image)
- Subline (8–18 words, will sit below the headline in smaller type)
- Brand bar (3–5 words max, e.g., show name or guest name)
- A 1-line "micro context" that explains who said it and where
  (e.g., "[GUEST], on the [SHOW NAME] podcast")
- An alt-text version (under 140 chars) for accessibility and SEO

Style direction: [pick one: minimalist, bold sans-serif, newsroom,
vintage typewriter, neon, brutalist].

Constraint: never invent a quote. Use the exact words from the input.

Example output:

Headline: “Consistency is a cope.” Subline: “Posting more rarely fixes the wrong problem.” Brand bar: “The Refresh - Ep. 142” Micro context: “[GUEST], on The Refresh podcast” Alt-text: “Quote card: ‘Consistency is a cope. Posting more rarely fixes the wrong problem.’ - [GUEST] on The Refresh, Ep. 142.” Style: brutalist

Pro tips:

  • Static quote cards underperform video by reach, but outperform on saves. Saves are the most valuable signal in 2026 because they trigger follower growth in recommendations.
  • Always write alt-text. It helps with accessibility, image SEO, and AI-overview citation.

Prompt 13 - The Title-Tester

Purpose: Generate 10 candidate titles per clip and predict which will win. A/B titles are the highest-leverage split test in your workflow.

Full prompt:

I have 10 podcast clips. For each clip, generate 10 working titles
and rank them.

For each title, score 0–10 on:
- Clarity (does a stranger know what they'll learn?)
- Curiosity (do they have to click to find out?)
- Specificity (is there a number, name, or unusual word?)
- Length (penalize anything over 70 chars)
- Clickbait risk (penalize if it oversells or lies)

Output a table. Total = 50 rows.

Clips:
1. [PASTE CLIP 1 QUOTE]
2. [PASTE CLIP 2 QUOTE]
…
10. [PASTE CLIP 10 QUOTE]

Example output (1 of 10 rows):

#TitleClarityCuriositySpecificityLengthClickbaitTotal
1”The 4 AM rule that saved my marriage”9797032
2”He almost quit. Then this happened.”6958-226
3”What 4 AM taught me about love”7667026

Pro tips:

  • Pick the highest-total title. If two are tied, pick the one with the highest curiosity score.
  • Use the title in three places: the YouTube title, the LinkedIn post, and the email subject line. They reinforce each other.

Prompt 14 - The “Mute-the-Sound” Caption Test

Purpose: 80%+ of social video is watched on mute. If your caption fails the “no sound” test, the clip dies. This prompt audits your caption and rewrites it for mute-watchers.

Full prompt:

I'm posting this podcast clip on TikTok, Reels, and YouTube Shorts.
85% of viewers will watch on mute. Audit and rewrite the captions.

Input clip transcript (verbatim):
"[PASTE TRANSCRIPT]"

Deliver:

1. The current caption: cut or keep?
2. A "mute-friendly" rewrite:
   - Each line: 4–8 words
   - Use sentence-case, not ALL CAPS
   - Highlight 1 power word per line in BOLD
   - Use 1 emoji per 3 lines max
   - Sync punctuation to natural speech pauses
3. A "Power Words Audit" - list 3 words from the rewrite that earn
   the most thumb-pause and why.
4. A "Sound-on Version" - same content, but with 1 sound effect call
   (e.g., "[record scratch]") for sound-on viewers.
5. A 1-sentence "caption-first" rule that should govern every clip
   you post.

Example output (abridged):

Current caption: keep, with edits. Mute-friendly rewrite:

  • “He almost quit last year.”
  • “$0 in the bank. Two kids.”
  • “Then he sent one email.”
  • “Here’s what happened next.” Power words: “$0,” “one email,” “next.” Sound-on version: same as above + “[whoosh]” before “Here’s what happened next.” Caption-first rule: “If you covered the screen, would the caption still make sense? If not, rewrite it.”

Pro tips:

  • CapCut, Submagic, Captions.com, and OpusClip all auto-generate captions, but none of them apply your tone. This is where ChatGPT earns its keep.
  • Use the “one bold word per line” rule from the Submagic playbook. It lifts average watch time by 4–8% in my experience.

SECTION 4 - Caption and headline prompts (Prompts 15–18)

You’re producing clips. Now you need the metadata: titles, descriptions, hashtags, pinned comments. These four prompts handle that.

Prompt 15 - The YouTube-Shorts Metadata Pack

Purpose: YouTube Shorts has its own algorithm. Title length, description phrasing, hashtag count, and pinned comment all matter differently than on TikTok or Reels.

Full prompt:

Generate the full YouTube Shorts metadata for this podcast clip.

Clip summary: [PASTE 1-SENTENCE SUMMARY]
Spoken quote: [PASTE THE QUOTE]
Guest: [NAME]
Show: [NAME]

Deliver:

1. **Title** (max 70 chars, must include 1 number or named entity,
   no clickbait)
2. **Description** (max 150 chars, must include the show name, guest
   name, and 1 natural keyword phrase)
3. **Hashtags** (exactly 3, no #Shorts, no #viral, no #fyp)
4. **Pinned comment** (1 sentence, asks a question that triggers
   replies, no link)
5. **First 2 lines of the show-note** (under 60 words each, optimized
   for YouTube search suggestions)
6. **End-screen CTA line** (under 20 words, will be read by the host
   in the last 5 seconds of the clip)

Example output:

Title: “[GUEST] lost $14,723 in 90 days. Here’s the fix.” Description: “On Ep. 142 of [SHOW], [GUEST] breaks down the 90-day cash crunch that almost killed the business. Podcast clip strategy.” Hashtags: #podcastclips #contentstrategy #podcastgrowth Pinned comment: “What’s the worst 90-day stretch you’ve ever had? Reply with the number.” Show notes line 1: “Why 90% of creators confuse effort with progress - and what to do instead.” End-screen CTA: “If this saved you a quarter, follow for the next one.”

Pro tips:

  • YouTube’s algorithm weights replies more than likes. The pinned comment is your cheat code.
  • 3 hashtags, not 30. YouTube’s own creator docs and Riverside’s December 2025 trends roundup both confirm that 3–5 targeted hashtags outperform 20 generic ones.

Prompt 16 - The TikTok-Specific Caption Pack

Purpose: TikTok is its own world. Hashtag count, sound choices, and on-screen text all behave differently than on YouTube or Reels.

Full prompt:

Build the TikTok-specific posting metadata for this podcast clip.

Clip context: [PASTE]
Guest: [NAME]
Show: [NAME]

Deliver:

1. **TikTok caption** (max 150 chars, must contain 1 question + 1
   keyword)
2. **Hashtag set** (5–8, mixing 1 broad + 2 mid + 2 niche + 1 branded
   show tag)
3. **On-screen text overlay** for the first 3 seconds (under 30 chars)
4. **Sound recommendation** - describe a 2026-trending audio style
   (e.g., "soft piano loop," "viral dialogue from a 2025 series"),
   but do not invent a specific copyrighted track name
5. **"Stitch-bait" line** - a 1-sentence line that ends the clip
   and invites duets/stitches without sounding desperate
6. **Comment seed** - 1 self-reply from the host account in the first
   10 minutes, written to prompt viewer replies

Example output:

Caption: “We lost $14k in 90 days. The fix was $9. Want it? 👇” Hashtags: #podcast #smallbusiness #contentcreator #founderstory #startup #thepodcastshow On-screen text: “How we lost $14k” Sound: “Soft lo-fi loop, 70–80 BPM” Stitch-bait: “Tell me your worst 90-day stretch - I’ll send the spreadsheet to the best one.” Comment seed (host): “Genuine question - what’s the cheapest fix that saved your business?”

Pro tips:

  • TikTok’s 2026 discovery surface rewards comment density in the first hour. Pin the host’s self-reply and reply to every comment for the first 60 minutes.
  • 5–8 hashtags, not 30. The data is consistent across Hootsuite, Buffer, and Riverside’s December 2025 report: niche beats volume.

Prompt 17 - The LinkedIn Authority Post

Purpose: LinkedIn is a different beast in 2026 - Hootsuite calls it the platform’s “creative era” (Hootsuite 2026). Text posts now compete with carousels and video. This prompt writes a personal, professional post that won’t sound like every other LinkedIn influencer.

Full prompt:

Write a LinkedIn post based on this podcast clip.

Source moment: [PASTE]
Guest: [NAME], [TITLE]
Show: [NAME]

Constraints:
- 900–1,300 characters (LinkedIn sweet spot)
- First line is a 1-sentence hook that creates an "information gap"
- Use 1 short paragraph (1–2 sentences) per idea
- No bullet lists (LinkedIn's algorithm de-prioritizes them in 2026)
- End with a single, open question
- Use 0–2 emoji total
- No hashtags in the body. Add 3–5 hashtags on a new line at the end.
- Voice: first-person, reflective, not salesy
- Include 1 specific number or named entity from the clip

Also deliver:
- A "carousel version" outline (7 slides, 1 idea per slide)
- A 90-character "document title" if I turn it into a PDF

Example output (abridged):

“I lost $14,723 in 90 days.

Then a $9 tool saved the business. Here’s the line item I missed for two months.

…

What’s the cheapest fix that ever saved your business? Curious which one shows up most.”

Pro tips:

  • Post Tuesday–Thursday between 8 AM and 10 AM in your audience’s timezone.
  • The “creative era” insight from Hootsuite means carousel posts now drive 2× reach over text-only. Use the carousel outline the prompt returns.

Prompt 18 - The Newsletter Headline Pack

Purpose: Email is still the highest-ROI channel for podcasters. The subject line decides if anyone sees the clip.

Full prompt:

Write 6 email subject lines and 3 preview-text options for an email
that will feature this podcast clip.

Clip: [PASTE]
Show: [NAME]
Sender name: [NAME]
Audience: [AUDIENCE]

Constraints:
- Subject line max 60 chars
- Preview text max 90 chars
- 2 of 6 should be questions
- 2 of 6 should be statements
- 2 of 6 should be numbers
- 1 of 6 must use the "you/your" framing
- None can use ALL CAPS, no emoji in subject lines

Also deliver:
- A 1-sentence preview text that complements (does not repeat) the
  subject line
- A "from-line" suggestion (sender name format)
- A "send time" recommendation (day + hour in user's local time)
  for a [B2B / B2C / creator] audience

Example output (abridged):

Subject lines:

  1. “The $9 fix that saved a $0 business”
  2. “Did you make this same mistake?”
  3. “We lost $14,723 in 90 days”
  4. “Your cheapest growth hack”
  5. “What nobody tells you about month 3”
  6. “3 things I’d do differently at day 1” Preview text: “Spoiler: the cheapest line item is the one you forgot to track.” From line: “[NAME] from [SHOW]” Send time: Tuesday 9:30 AM ET (B2B)

Pro tips:

  • A/B test subject lines. Most ESPs (ConvertKit, Beehiiv, MailerLite) handle this automatically.
  • Preview text is the most-skipped A/B test in email. The same line that wins the subject line often loses as preview text. Always co-test.

SECTION 5 - Distribution and repurposing prompts (Prompts 19–22)

You have clips, captions, and metadata. Now distribute them. These four prompts automate the last mile.

Prompt 19 - The 14-Day Content Calendar Generator

Purpose: Turn 12 finished clips into a 14-day posting calendar. Pat Flynn’s “Editorial Calendar” framework, updated for short-form.

Full prompt:

I have 12 finished podcast clips with metadata. Build me a 14-day
content calendar.

Clips:
1. [TITLE] - primary platform: [P] - hook: [H]
…
12. [TITLE] - primary platform: [P] - hook: [H]

Deliver:

1. A 14-row table with: Day, Date, Time, Platform, Clip #, Format
   (video / static / thread / email), Caption excerpt (under 30 chars),
   Hashtags, Goal (reach / saves / clicks / replies)

2. Posting cadence rules:
   - 1 clip/day on TikTok
   - 1 clip/day on Reels
   - 4 clips/week on YouTube Shorts
   - 1 thread/week on X
   - 1 carousel/week on LinkedIn
   - 1 email/week

3. A "do not post on" list (the days of the week that historically
   underperform for this audience)

4. A "boost candidate" - the 1 clip most likely to perform that
   should get a paid boost ($25–$50) on day 3

5. A 1-sentence "kill rule" - the metric + threshold at which I
   should pull a clip early and replace it (e.g., "if watch-time
   share is below 30% after 6 hours")

Example output (abridged):

DayPlatformClipFormatGoal
MonTikTok#1VideoReach
MonReels#2VideoReach
TueYouTube Shorts#3VideoReach
TueXThread #1ThreadReplies
…
SunNewsletter#12EmailClicks

Boost candidate: Clip #4 (the $14,723 story) on Day 3, TikTok, $40. Kill rule: Pull any Reel that doesn’t hit 35% watch-time share by hour 8.

Pro tips:

  • Calendars beat willpower. The “Editorial Calendar” idea, popularized by Pat Flynn, still wins in 2026 - without it, clips go out randomly and underperform.
  • Stack 2–3 platforms per day, not 1. Cross-posting via Repurpose.io (or OpusClip’s built-in scheduler on its Pro plan) saves 4–6 hours per week.

Prompt 20 - The Cross-Platform Adaptor

Purpose: A clip made for TikTok rarely works on LinkedIn. This prompt takes one finished clip and writes 4 platform-specific versions, each with the right tone, length, and CTA.

Full prompt:

I have one finished podcast clip. Write 4 platform-specific versions.

Source clip (30 seconds, TikTok-native):
"[PASTE TRANSCRIPT + METADATA]"

Deliver, in this order:

1. **TikTok version** (30s, on-screen caption-style captions, 1
   trending sound cue, 5–8 hashtags, stitch-bait end)

2. **Reels version** (30s, polished captions, 1 CTA, 3–5 hashtags,
   end with "save this for later")

3. **YouTube Shorts version** (35s, slightly longer hook, optimized
   for search - include the keyword phrase in the first 3 seconds of
   spoken audio, 3 hashtags)

4. **LinkedIn version** (45s, professional caption, 0 hashtags in
   caption, 3–5 hashtags in the first comment, ends with a question
   for replies)

For each, return:
- The full spoken text (what the voiceover / host says)
- The on-screen text overlay (3–5 lines)
- The caption / first comment text
- The CTA

Constraint: keep the same quote. Adapt the wrapper, not the core.

Example output (LinkedIn, abridged):

Spoken text: “We lost $14,723 in 90 days. The cheapest line item was a $9 tool nobody told us to track.” On-screen text: “Lost $14,723 / Found a $9 fix” Caption: “The line item we missed for two months. Curious what yours is.” First comment: “#leadership #smallbusiness #founderlife #startup #business” CTA: “Reply with the cheapest fix that ever saved your business.”

Pro tips:

  • Most “cross-posting” tools just re-upload. They don’t adapt. This is where ChatGPT earns its keep.
  • Run the 4 versions on the same day at staggered times. You can post 2–4 versions of the same core clip without it being seen as spam if the wrappers differ.

Prompt 21 - The Repurpose.io Workflow Planner

Purpose: Repurpose.io is a 2026 workhorse for podcasters - it auto-pushes one video to many platforms. This prompt writes the actual Repurpose workflow (or the manual equivalent if you don’t have the tool).

Full prompt:

I publish a 60-minute podcast video on YouTube every Tuesday at 6 AM ET.
I want to automatically cross-post 8 vertical clips per episode to
TikTok, Reels, YouTube Shorts, and LinkedIn.

Design the full workflow. Include:

1. **Source** (YouTube full video, RSS, Dropbox, etc.)
2. **Trigger** (when does the workflow run - manual, scheduled, or
   "on new upload"?)
3. **Auto-clipping step** (OpusClip, Riverside Magic Clips, or
   Descript - pick one and say why)
4. **Human review step** (what I check before approval - captions,
   aspect ratio, hook quality)
5. **Approval** (Slack / email / native app)
6. **Distribution** (the order of platforms, with delay between
   each post)
7. **Failure handling** (what happens if a post fails, who gets
   notified, and how soon)
8. **Daily time cost** (estimated minutes of human work per day)
9. **Tool stack** (a clean diagram in text form: source → clip →
   caption → approve → publish)

End with the 3 things that will break this workflow first, and how
to prevent them.

Example output (abridged):

Tool stack: YouTube RSS → OpusClip auto-clips (12 candidates) → Notion review queue → ChatGPT caption polish → Repurpose.io → TikTok, Reels, Shorts, LinkedIn Daily time cost: 22 minutes per day after week 2. Will break first: (1) YouTube RSS hiccup, (2) OpusClip rate limit, (3) Repurpose.io token expiry. Mitigations: backup RSS polling, Pro plan upgrade, 90-day token calendar.

Pro tips:

  • Repurpose.io’s own docs confirm a 5-platform push in under 8 minutes of human time once tuned.
  • OpusClip’s Starter plan ($15/mo) includes 150 credits and is the right size for one weekly show. The Pro plan ($29/mo, $14.50 annual) adds AI B-roll, brand templates, and the social scheduler - a 2× ROI if you post daily.

Prompt 22 - The “1 Episode to 30 Clips” Engine

Purpose: The flagship prompt. Drop in one transcript, get back a full production plan for 30 short-form clips plus a 30-day calendar. Use this once a month and your pipeline is set.

Full prompt:

I'm giving you one full podcast episode transcript. I need you to
produce a complete 30-clip content plan.

Inputs:
- Transcript: [PASTE]
- Show: [NAME]
- Guest: [NAME]
- Audience: [AUDIENCE]
- Primary platforms: [TikTok / Reels / YouTube Shorts / LinkedIn / X]
- Posting cadence: [e.g., 2 clips/day across platforms]

Deliver:

1. **30 clip candidates** in priority order. For each:
   - Title (max 70 chars)
   - Spoken text (verbatim, 10–40 seconds)
   - On-screen caption version (under 90 chars)
   - Hook-Insight-CTA breakdown
   - Best platform (primary + secondary)
   - Suggested CTA
   - Clip score 0–10

2. **A 30-day posting calendar** with day, platform, clip #, time,
   and goal (reach/saves/clicks/replies).

3. **3 evergreen re-cut ideas** for clips that could be re-edited at
   90/180/365 days.

4. **A 4-week rollout plan** (week 1 = launch wave, week 2 = re-cut
   and remix, week 3 = testimonial/quote-card static wave, week 4 =
   summary compilation).

5. **A 1-paragraph "behind the scenes" caption** I can post on the
   show's Instagram story to invite the audience into the process.

Constraints:
- Never invent quotes
- Never attribute lines to the wrong speaker
- Surface at least 5 numerical claims, 5 emotional peaks, and 5
  contrarian takes across the 30 clips

Example output (abridged, first 3 of 30):

Clip #1 - Score 9.4 - “The 4 AM rule that saved my marriage” Clip #2 - Score 9.1 - “We lost $14,723 in 90 days” Clip #3 - Score 8.8 - “Consistency is a cope”

Pro tips:

  • Run this prompt once per episode. The calendar is the artifact - the clips are byproducts.
  • Pair with OpusClip’s calculator (visible on its pricing page) to estimate time saved.

Comparison table: prompt category vs. clip type vs. output

Prompt categoryClip type producedTypical lengthBest-fit platformTool to pair with
Pre-recording seed (1–3)Steered interview moments30–90s rawAll platformsChatGPT, guest research tools
Transcript mining (4–9)Raw clip candidates + score10–45sTikTok, Reels, ShortsDescript, Riverside, Otter, Tactiq, Fathom, Read AI
Quote rewrite (10–14)Polished clips + variants7–35sTikTok, ReelsCapcut, Submagic, Captions.com
Caption & headline (15–18)Metadata + thumbnailsn/aAll platformsCanva, Headliner
Distribution (19–22)Calendar + cross-post assetsn/aAll platformsOpusClip, Repurpose.io, Buffer, Hootsuite

How to read this: The arrows run left to right. Pre-seed (Prompts 1–3) shapes the raw material. Mining (4–9) finds the moments. Rewrite (10–14) compresses them. Caption (15–18) packages them. Distribution (19–22) sends them out. The right-side tools are optional - the prompts stand alone, but each has a recommended companion. OpusClip, for example, runs the auto-clipping step; ChatGPT does the “Clip Score” prioritization in parallel and the two correlate about 80% of the time.

People Also Ask: 10 questions, answer-first

Q1: What are the best ChatGPT prompts for podcast clip-worthy quotes? The best prompts pair a transcript with a scoring rubric. The Prompt 4 (“25-Clip Candidate Sweep”) and Prompt 9 (“Hook-Insight-CTA Reframer”) are the two I rely on most. They combine a clear filter set (contrarian, numerical, emotional, named framework) with a 0–10 score you can sort by.

Q2: How do I extract quotes from a podcast using AI? Three steps. (1) Transcribe the audio in Descript, Riverside, Otter, Tactiq, Read AI, or Fathom. (2) Paste the transcript into ChatGPT and ask for clip candidates with a score. (3) Use OpusClip or Riverside’s Magic Clips to render the clips with auto-captions. The OpusClip pricing page confirms its Pro plan ($29/mo, billed annually at $14.50) includes AI clipping, animated captions, AI B-roll, and direct posting to YouTube Shorts, TikTok, and Instagram.

Q3: How many clips should I make from one podcast episode? Industry data points to 8–30 clips per 60-minute episode, with 15 being the sweet spot for solo creators. Hootsuite’s 2026 trends report calls out “the micro-drama trend” - short-form series and content clipping - as a $7.8B market (Deloitte, 2026). More clips, distributed consistently, beat fewer clips posted sporadically.

Q4: What is the Hook-Insight-CTA framework? A 3-part clip structure popularized by Think Media: (1) Hook (3–7 words to earn 3 more seconds of watch time), (2) Insight (15–25 seconds of compressed value), (3) CTA (a single command - comment, save, share, follow). It’s the de facto template inside OpusClip, Riverside Magic Clips, Submagic, and CapCut.

Q5: Can ChatGPT write YouTube Shorts titles? Yes. Use Prompt 15 to generate title + description + hashtags + pinned comment in one pass. The 70-character title limit, the 3-hashtag rule, and the “first 2 lines of show notes” structure are all YouTube-Search-optimized for 2026.

Q6: How do I use ChatGPT with OpusClip? Use ChatGPT to prioritize and rewrite clip candidates (Prompts 4, 9, 14). Use OpusClip to auto-render them with virality scores, AI reframing, animated captions, and direct publishing. The OpusClip calculator on its pricing page estimates that 150 credits/month on the Starter plan yields ~150 clips, ~100 hours saved per week, and ~351K extra views - numbers to sanity-check against your own channel.

Q7: What’s the best free way to turn a podcast into short clips? Stack ChatGPT (free tier works) + OpusClip free plan (60 credits/month, up to 1080p, watermark) + CapCut’s free auto-captions + Repurpose.io’s free tier (1 source, 1 destination). The watermark is removable in CapCut. This stack costs $0 and produces publishable clips - verified on OpusClip’s pricing page as of June 2026.

Q8: How long should a podcast clip be in 2026? 7–35 seconds is the new sweet spot. YouTube Shorts rewards 30–35s for completion. TikTok rewards 15–25s for share. LinkedIn rewards 45–60s for comment density. The “Quoted Card” static format does best at 8–12 words on a 1080×1350 image. Riverside’s December 2025 report confirms completion rate drops sharply after 35 seconds in vertical.

Q9: Do podcast clips grow followers? Yes - and the conversion is measurable. 44% of weekly podcast listeners have purchased a product after hearing it on a podcast (Edison Research / Riverside, 2025). Clips are the top of the same funnel. A single well-distributed clip can pull 1,000–10,000 targeted impressions and convert 1–3% to follows, per Hootsuite’s 2026 trends report.

Q10: How do I write a viral podcast clip caption? Lead with a number, a “you/your,” a sharp verb, or a question - never with “Did you know.” Use 1 bold word per caption line. Cap the caption at 150 characters. End with a single CTA. Pair with a hook that passes the “mute-the-sound” test. (This is the full system behind Prompt 10 and Prompt 14.)

The 7-day workflow: 1 episode to 30 clips

Here’s the exact week I’d run if I were launching a new show today. I built it from the prompts above and from how OpusClip, Riverside, Descript, Repurpose.io, and Buffer actually behave in June 2026.

Day 1 (Mon) - Record + seed. Run Prompt 1 and Prompt 2 the day before. Use Prompt 3 for the cold open. Record 60–90 minutes.

Day 2 (Tue) - Transcribe + mine. Upload the raw file to Descript or Riverside for transcript. Paste the cleaned transcript into ChatGPT and run Prompt 4. You’ll get 25 clip candidates with scores. Pick the top 12.

Day 3 (Wed) - Render. Push the 12 candidates through OpusClip (or Riverside Magic Clips). Pick the top 8 to render at 1080×1920 with auto-captions.

Day 4 (Thu) - Rewrite + caption polish. Run Prompt 9 on each of the 8 clips. Run Prompt 14 to audit captions. Run Prompt 10 to write 5 cold-open options per clip. Pick the best.

Day 5 (Fri) - Metadata. Run Prompt 15 (YouTube Shorts), Prompt 16 (TikTok), Prompt 17 (LinkedIn), and Prompt 18 (email) on the strongest 4 clips. Build a static “Quoted Card” with Prompt 12 for one numerical moment (Prompt 7’s output).

Day 6 (Sat) - Calendar + adapt. Run Prompt 19 to build the 14-day calendar. Run Prompt 20 to adapt the top 4 clips to all four platforms. Push the 4 cross-platform versions via Repurpose.io.

Day 7 (Sun) - Newsletter + recycle. Send the email using Prompt 18’s subject lines. Pull one tweet thread from Prompt 11. Pin your favorite clip on YouTube. Plan next week’s pre-seed prompts. Total active time: ~3 hours.

By the end of week 1: 8 polished clips, 4 cross-platform versions, 1 newsletter, 1 thread, 1 carousel, 1 quoted card, 14-day calendar, 30-day forecast. All from one 60-minute episode.

Common mistakes to avoid

I’ve made all of these. You don’t have to.

Mistake 1 - Skipping the transcript cleanup. Descript and Riverside give you a near-perfect transcript, but “um”s, false starts, and speaker mislabels leak in. A 90-second cleanup pass before Prompt 4 changes the output quality dramatically. Prompt 4 returns garbage if the transcript is garbage.

Mistake 2 - Trusting the AI score blindly. The “Clip Score” in Prompt 4 is a heuristic. Your gut is a heuristic too. Use both. If ChatGPT scores a clip 8.5 and you felt nothing when it played, kill it. The clip that makes you rewind is the clip that travels.

Mistake 3 - Same hook on every clip. The fastest way to flatten your reach is to use “Here’s why…” or “Most people don’t know…” as the cold open on every post. Vary the hook. Prompt 10’s job is to fight this. So does the A/B test in Prompt 13.

Mistake 4 - Posting and praying. A clip that underperforms in its first 6 hours can be revived with a new thumbnail (Prompt 12), a new caption (Prompt 14), or a new hook (Prompt 10). Hootsuite’s 2026 report (hootsuite.com/research/social-trends) calls this the “fastvertising” trend - fast iteration beats big launches.

Mistake 5 - Ignoring the email list. The clip-first world forgets that email is the only algorithm-free channel you own. Use Prompt 18 weekly. Even a 200-person list will outearn a 20,000-follower TikTok account over 12 months.

Mistake 6 - Treating ChatGPT like a search engine. ChatGPT is a thought partner. The prompts in this guide are prompts because they have a voice and a constraint. If you strip the constraints, you strip the magic. Copy the prompts. Then adapt them to your show.

Mistake 7 - Using the wrong tool for the wrong step. ChatGPT does the thinking. OpusClip and Riverside do the cutting. Descript does the transcript. CapCut and Submagic do the captions. Repurpose.io does the distribution. Headliner does the audiogram. Buffer does the static cards. Trying to make one tool do all of this is where most creators stall.

Final word

The math hasn’t changed since podcasting started: you record, you distribute, you grow. What changed in 2026 is the amplification curve. With 619.2 million listeners and 50.6% of shows on YouTube (Riverside, Dec 2025), the long tail is longer than ever and the long-form content is more discoverable than ever. Hootsuite’s 2026 report calls it the search-first trend: Google now indexes public Instagram content and short-form video from every major platform. A clip you post on Tuesday can show up in a search result for years.

The 22 prompts above are your refinery. They turn a 4,000-word transcript into a 14-day content calendar, a 30-day forecast, and a re-usable engine. They cost you ChatGPT, OpusClip, and 3 hours per episode. The ROI is asymmetric: one good clip can carry your show for a quarter.

Pick three prompts to start. I’d begin with Prompt 4, Prompt 9, and Prompt 19. They cover the full pipeline in miniature. Run them on your next episode, score the output against your last month, and iterate.

The audio you already recorded is full of lines that could be working for you right now. ChatGPT is the fastest way I know to find them.