Muse for Creators and Podcasters: Episode Planning, Show Notes, and One Recording Turned Into a Week of Posts
A creator's working setup for Muse: plan the episode and research the guest before you record, turn the transcript into show notes with real timestamps, and repurpose one recording into clips, a newsletter, and a week of posts, with every quote checked against the transcript.
By Professor Muse13 min read
A weekly show is two jobs wearing one name. The first is the recording, and it is the part people sign up for. The second is everything after it: the episode plan, the guest research, the show notes, the clips, the newsletter, the posts that keep the episode alive after launch day. The second job is longer than the first, and it decides whether a show grows or stalls, because a recording nobody can find, skim, or share does not compound.
This lesson sets up three jobs that run the same way every week. Muse plans the episode and researches the guest before you record, turns the finished transcript into show notes with timestamps a listener can use, and repurposes the one recording into a batch of assets that publish over the next week or two. You keep the editorial calls, the guest relationship, and the final read. Muse does the drafting, the formatting, and the scheduling list.
One rule holds the whole setup together: the transcript is the source document, and everything written about the episode must trace back to it. A quote, a timestamp, a claim about what the guest said, all of it comes from the transcript or it does not go out. That keeps a fast workflow honest.
The short version
- Plan before you record: one episode plan with a run of show, three questions that only this guest can answer, and a research brief with sources attached.
- The transcript is the source document for everything after the recording. Show notes, clips, quotes, and the newsletter all trace back to it.
- Show notes need a hook, a short summary, three to seven takeaways, timestamps that start at 00:00, and one call to action. Timestamps come from the final edit, never the raw file.
- Repurpose, do not cross-post: each asset gets its own hook and length for its channel, and the batch spreads over one to two weeks instead of landing all on launch day.
- The failure mode to watch for is a quote attributed to the wrong speaker, or a quote that was never said. The fix is quotes only with timestamps, verified against the transcript.
The Setup: The Thursday Interview
Call the whole arrangement the Thursday Interview. I run a weekly interview show, one 45-minute conversation, recorded on Thursdays and published on Fridays. The guest is usually a founder or a builder with one specific thing they know well. The episode goes out as audio on podcast apps and as video on YouTube, and the same recording feeds everything else: show notes on the episode page, three to five short clips, a newsletter summary, and a handful of social posts across the following week.
The setup has four parts, and they never change. A standing instruction from our memory guide sets the house rules: never invent a quote, never attribute a line to a speaker without a timestamp, never publish a guest’s name, title, or link without checking it against how the guest writes it, and never send anything to a guest without a human read. The episode plan lives in one document per episode, with the run of show, the research brief, and the questions in one place. The transcript is the source file for every written asset, and the asset map, the list of everything one episode becomes, is written before any asset is drafted. The weekly run is a scheduled routine, the same shape as a Friday expense review: the recording lands Thursday, the show notes and clips are drafted Friday, and the posts go out across the next week.
The same planning discipline carries into the marketing side of the show:
Prompt 1: Plan the Episode and Research the Guest
Run this a week before you record, while there is still time to change the plan. The prompt asks for research first and questions second, because a question worth asking comes from something specific the guest has actually done or said, not from a generic list. Give Muse the sources: the guest’s site, their two most recent interviews or talks, and anything they have written on the topic.
Plan episode [number] of the Thursday Interview with [guest name, title, company]. Research the guest from these sources: [links to their site, recent interviews, talks, or writing]. Write a one-page episode plan with four parts. (1) A research brief: what the guest is known for, the one project or decision that defines them, and the two most interesting things they have said in public, each with the source it came from. (2) A run of show with timed blocks for a 45-minute conversation: cold open, guest background, the core topic, a practical section, and the close. (3) Eight questions, in the order I should ask them, with the three that only this guest can answer marked. No generic questions a guest could answer about anyone. (4) A short list of topics to avoid, based on what the guest has declined to discuss in public, and anything in the research that looks uncertain or secondhand, flagged as unverified. Do not invent biographical details, dates, or quotes. If a fact is not in the sources, say so.
Sources: [paste the links]. Guest’s own wording for their name and title: [paste it if you have it].
Read the research brief against the sources before you trust it. The brief is where the episode is won or lost, and a wrong fact in the brief becomes a wrong question on the recording, where it is much harder to fix. Check the guest’s name and title the way you would check a spelling in a contract: against how the guest writes it themselves, not against a third-party bio. If the brief flags something as unverified, treat it as a question to ask.
Two details matter at the planning step. Record the run of show as named blocks, because named blocks become the chapters in your show notes later. An episode planned as “background, the problem, the framework, mistakes, close” already has its timestamps half written. Ask the guest for their preferred bio and links in the booking confirmation, the same way a booking form collects the details that end up in the invite. The bio you were given beats the bio you found.
Prompt 2: Write the Show Notes From the Transcript
Run this on the transcript of the final edit, after the episode is cut, never on the raw recording. Ads, intros, and cut sections shift every timestamp after them, and show notes built from the raw file point listeners at the wrong minute. A transcript from the final edit is the only one whose times are true.
Write the show notes for episode [number], [episode title], from the attached transcript of the final edit. Follow this structure. (1) A hook: one or two sentences that say why this episode is worth an hour, with the guest’s name and the core topic in the first line. (2) A summary of 100 to 200 words, with the main keyword near the start. (3) Three to seven specific takeaways, each one sentence. (4) Timestamps as chapters: start at 00:00, three to six chapters per hour, each title under five words, one per line, in time order, with each time checked against the transcript. (5) A guest bio under 75 words, using the wording the guest gave us, plus every link mentioned in the episode. (6) One call to action, not five. Rules: every quote must come from the transcript with its timestamp, and no quote may be paraphrased and presented as a direct quote. Every link must be one that was actually mentioned or that the guest provided. Keep the whole note under 4,000 characters so it fits the RSS feed. If the transcript is missing a speaker label for a quote, say so instead of guessing who said it.
Transcript: [attach the transcript of the final edit]. Guest bio and links as provided: [paste them]. Episode links: [the episode page, the full transcript link].
Read the timestamps against the final edit before you publish. Podcast apps turn a well-formed timestamp list into tappable chapters, and a list that starts anywhere but 00:00, runs more than six chapters an hour, or carries titles longer than five words breaks that conversion. A listener who taps a chapter and lands in the wrong segment stops trusting the notes, and the notes are the thing that decides whether a stranger presses play.
The hook deserves the same care. Podcast apps cut the description short, so the first line has to work alone. A hook that says what the guest delivers, in their words, beats a hook that says the episode is interesting. If the first line does not survive on its own, rewrite it until it does.
Prompt 3: Repurpose One Recording Into a Week of Posts
Run this after the show notes are done, because the show notes already hold the decisions this prompt needs: the best moments, the takeaways, and the timestamps. Repurposing starts with an asset map, a list of everything one episode becomes. Write it before drafting any asset. A practical default is three to five short clips, show notes with timestamps, one newsletter summary, and three to seven social posts, published over one to two weeks.
Build the repurposing batch for episode [number] from the attached transcript and the show notes below. (1) Write an asset map: the full episode, three to five short clips, one newsletter summary, and three to seven social posts, each with the channel it is for and the day it publishes. (2) For each clip, give the start and end timestamp, the one idea it makes, and a caption with the hook in the first two or three seconds. Pick moments that make sense without context: a sharp opinion, a surprising answer, a short story with a payoff, or a practical tip. Cut each clip to 20 to 45 seconds for vertical formats, captioned, since many people watch with the sound off. (3) Write the newsletter summary: the three most useful points and one link to the episode. (4) Write the social posts: one quote post per clip, each quote verbatim from the transcript with its timestamp, and one longer post that makes a single point from the episode. Rules: adapt each asset to its channel, do not paste the same copy everywhere. Every quote must be verbatim with a timestamp. If a moment needs context to make sense, it is not a clip.
Transcript: [attach the transcript of the final edit]. Show notes: [paste them]. Publishing window: [the next two weeks].
Spread the batch out. The full episode, the show notes, and the newsletter go out on launch day. The clips and the quote posts go out across the following days, one or two a day, and a couple of clips are held back to promote the episode again later. Publishing everything on the same day gives each asset no room to perform, and it burns the whole episode in one afternoon. The point of the batch is that one conversation keeps working for weeks.
The follow-ups around the episode follow the same cadence. Our meetings lesson covers the shape: a thank-you note to the guest with their clips attached and a short line on how to share them, sent the same day the episode goes live, and the guest’s approval sought before anything sensitive goes out. A guest who gets their clips with a usable caption shares them. A guest who has to ask where the episode is does not.
| Asset | When | What it answers |
|---|---|---|
| Episode plan and research brief | A week before recording | What is this episode about, and what can only this guest answer |
| Show notes with chapters | After the final edit | Is this episode worth an hour, and where do I jump in |
| Clips, 20 to 45 seconds, captioned | Launch day plus the following week | What is the one moment worth watching |
| Newsletter summary | Launch day | What were the three most useful points, and where do I listen |
| Social posts and quote cards | Spread over one to two weeks | What did the guest say, in their words |
What Can Go Wrong: The Quote Nobody Said
Here is the failure mode you will meet first. The show notes go out, the clips go out, and a listener messages you: the guest never said that line. The cause is a quote Muse assembled from two nearby sentences, or a line from the host attributed to the guest, or a paraphrase dressed up in quotation marks. The transcript was right there, and nobody checked the quote against it, because the quote read well and the workflow moved fast.
The fix has three parts, and all three are already in the prompts above. First, Prompt 2 forbids paraphrase as quotation: a quote is verbatim from the transcript or it is not a quote. Second, every quote carries its timestamp, so any quote can be checked in seconds by opening the transcript at that minute. Third, speaker labels are never guessed: if the transcript does not say who spoke, the draft says so instead of assigning the line to the person it sounds like. A quote with a timestamp and a named speaker is a fact. A quote without either is a risk, and it goes out under your name.
Timestamps need the same caution. A timestamp list built from the raw recording instead of the final edit is wrong in every entry after the first cut, and it is wrong in a way that looks right until a listener taps a chapter. The standing rule handles it: timestamps come from the final edit, checked against the transcript of that edit, and the check takes five minutes while the episode is still open. Our failure modes guide covers this pattern in general terms: a fluent output that reads cleanly gets less checking than a rough one, and the checking it skips is exactly the checking that would have caught the error. A quote check and a timestamp check are both fast, so run them every week, even when the notes looked clean.
Start With One Episode
Do not build the whole Thursday Interview this week. Take your last episode and run Prompt 2 on the transcript of its final edit. Read the hook on its own, check the timestamps against the edit, and verify every quote against the transcript at its timestamp. Then run Prompt 3 and draft the batch, and hold two clips back for next week.
Run it that way and the recording becomes the start of the week, not the end of it. One conversation keeps working long after the microphones are off.
The repurposing sequence in this lesson, transcribe the recording, pick the strongest moments, write an asset map, adapt each asset to its channel, and schedule the batch over one to two weeks, follows the repurposing guide published by Recast Studio, which lists the same default set of three to five short clips, show notes with timestamps, a newsletter summary, and three to seven social posts. The chapter rules, start at 00:00, three to six chapters per hour, titles under five words, and timestamps from the final edit, follow the show notes guide published by ParrotNotes, which checks them against the chapter rules Apple Podcasts and Spotify publish.
