Verified against ChatGPT · 2026-08-14
Turn a podcast episode transcript into show notes that actually get found in search, not just a summary paragraph
Produces structured show notes with timestamped chapters, a searchable summary, and pull-quotes built from what listeners actually say they came for, instead of a generic three-sentence recap.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Write show notes for the podcast episode below. Most show notes are a throwaway three-sentence summary that describes the episode without ever using the words a listener would actually search for — I want notes that function as a real landing page for this episode.
EPISODE TRANSCRIPT OR DETAILED SUMMARY
55-minute conversation with a sleep researcher about why consistent wake times matter more than total hours slept, including a mentioned 2023 study and a personal experiment the host tried.
GUEST NAME AND CREDENTIAL (if applicable)
Dr. Elena Kowalski, sleep researcher at a university lab, author of one book on circadian rhythm.
SPECIFIC TERMS OR QUESTIONS LISTENERS SEARCH FOR IN THIS TOPIC AREA
'why do I wake up tired even after 8 hours', 'consistent wake time vs sleep hours'.
RULES
Write an opening summary of 2-3 sentences that states the specific claims or topics covered, using the actual terms a searcher would type, not vague description ('we talk about productivity' is wrong; 'why time-blocking fails for people with unpredictable schedules' is right). Build a timestamped chapter list from the actual structure of the transcript, with each chapter title describing what's specifically discussed there, not a generic label like 'Discussion continues.' Pull 2-3 direct quotes from the transcript that are specific and quotable on their own, attributed correctly, not paraphrased. List any specific resource, book, tool, or study mentioned by name in the episode, since those exact-match terms are often what brings search traffic to an episode page. Do not editorialize or add promotional language not grounded in what was actually said.
WHAT NOT TO DO
Do not write a summary so generic it could describe half the episodes in the show's catalog. Do not invent a resource or quote that wasn't actually in the transcript — flag it as 'not found in transcript' rather than guessing if something referenced isn't clearly identifiable.
OUTPUT FORMAT
1. Episode summary (2-3 sentences).
2. Timestamped chapter list.
3. 2-3 pull-quotes with speaker attribution.
4. Resources/tools/studies mentioned, by name.
5. A suggested episode title distinct from the summary, if the current working title is generic.Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Show notes function as a web page that search engines can actually index, unlike the audio itself, which means the specific words used in the summary directly determine whether the episode surfaces for a relevant search — a vague summary like 'we talk about sleep' shares almost no vocabulary overlap with an actual search query like 'why do I wake up tired after 8 hours,' while a summary written to include that specific phrasing has a real chance of matching it. GPT-5.1's default instinct on 'write show notes' is to produce a polished but generic recap, because that's the dominant shape of the training examples for the phrase 'show notes' — this prompt overrides that by explicitly requiring the summary to use searcher vocabulary rather than description, and by requiring named resources and direct quotes, which are the two elements most likely to contain the specific proper nouns and phrases a search engine can match against. The instruction to flag an uncertain resource as 'not found in transcript' rather than guess directly addresses a known model failure mode — filling a plausible-sounding gap with an invented but wrong detail — which is especially damaging in show notes because a wrong book title or study citation is exactly the kind of error a listener or the guest themselves is likely to notice and flag publicly. Requiring chapter titles that describe what's specifically discussed, rather than generic section labels, also serves a secondary function beyond search: podcast platforms increasingly surface chapter markers directly in the player UI, so a vague chapter title is a missed opportunity twice over, both for search indexing and for in-app navigation.
What you get back
Summary: Sleep researcher Dr. Elena Kowalski explains why waking up tired after a full 8 hours often comes down to inconsistent wake times rather than total sleep duration, and walks through a 2023 study showing circadian misalignment's effect on perceived rest. [12:40] The wake-time-versus-hours distinction, explained. [24:15] The 2023 study and what it actually measured. [38:02] Host's own two-week wake-time experiment and what changed. Quote: 'Your body doesn't care how many hours you got if it never knows what time morning is supposed to be.' — Dr. Elena Kowalski.
Verified against
ChatGPT GPT-5.1 · 2026-08-14
Changelog
- 2026-08-14 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
