Podcast Episode Script is a Claude AI skill — a full, structured episode script calibrated to your show format, guest background, and audience expectations. Works inside Claude.
You spend two hours preparing for a guest interview. You read their LinkedIn, skim a couple of old podcast appearances, and draft a rough list of questions in a notes app. Then you ask Claude to turn it into a proper script. It comes back with an intro that could belong to any show, opener questions so broad the guest could answer them in their sleep, and a closing segment that sounds like it was lifted from a journalism 101 handbook. You rewrite most of it yourself. The AI saved you nothing.
The problem isn't Claude. It's that Claude had nothing to work with. No sense of your show's voice. No knowledge of what this specific guest has said before — and, crucially, what they haven't. No awareness of the format your audience actually responds to, or what's been covered to death on every other show in your niche. You handed it a blank canvas and expected a portrait.
A podcast script isn't a writing problem. It's a research problem with a writing problem attached. The Podcast Episode Script skill solves both — by doing the research first.
What Generic Scripting Actually Produces
Ask vanilla Claude for a podcast script and it will give you structure. You'll get an intro block, a guest bio read, a handful of questions grouped loosely by theme, and a sign-off. It looks like a podcast script. If you squint at it, it even feels like one. Then you try to record with it and realise the questions are all things the guest has been asked forty times. The opener bio is a recitation of their LinkedIn summary. The "insightful" follow-up questions are the exact softballs every host defaults to when they haven't done the work.
That's not Claude being bad at writing. That's what happens when a script is built without context. The guest's real story — the pivot they haven't talked about, the opinion that runs counter to received wisdom in their industry, the specific tension between their background and their current work — none of that is in the output. Because none of it was in the input.
A podcast script written without live research into the guest and current format trends asks the same questions in the same order as every other episode your listener already skipped.
Listeners are not patient. They know within the first three minutes whether this episode is going to give them something they haven't heard. A generic opener — "So tell us how you got started" — is a signal that you didn't prepare, and they'll act accordingly.
Why Live Research Changes the Script
Before the Podcast Episode Script skill writes a single question, it researches. It looks at what your guest has said publicly — what positions they've taken, what they repeat in every interview, and where the gaps are. It maps current format conventions for your show type, because what works in a founder interview format is structurally different from what works in a solo narrative show or a debate-style episode. These aren't interchangeable.
Static scripting — giving Claude a name and a topic and asking it to write — produces output anchored to training data. The format trends it knows are the ones that were circulating when it was trained. Guest research is whatever happens to be in its parametric memory. For a well-known figure, that might be something. For most guests on most podcasts, it's close to nothing.
The question that makes a guest lean forward is never the obvious one — it's the one that only exists if you've done the research.
Live research means the skill knows what your guest said in their last three public appearances, which means it can write the question that goes somewhere new. It knows what format patterns are currently driving engagement in your category, which means the pacing and structure of the script reflects what's actually working — not what worked two years ago.
What the Podcast Episode Script Skill Does
The skill asks three calibrating questions before it generates anything. Your show format, your guest's name and context, and the angle you want this episode to take. Those three answers transform the output from a generic interview template into something that fits your specific show with your specific guest on this specific topic.
Generic Script vs NovaKit Script
Same guest. Same topic. The difference is what went into the script before a single line was written.
Q1: "So tell us a bit about your background and how you got to where you are today."
Q2: "What's the biggest mistake you see early-stage founders making?"
Q3: "What does the future of SaaS look like in your view?"
Q1: "You've said publicly that most SaaS growth advice is written for companies that have already won. What's the version of that advice nobody's publishing?"
Q2: "In your last interview you mentioned [specific position]. Six months on — do you still hold that, or has something shifted?"
Q3: "If you had to disagree with the most widely-held belief in your space, what would it be?"
The difference isn't tone. It's depth of preparation visible in the questions themselves. The first version signals that the host hasn't listened to any prior appearances. The second signals that they have — and that this episode is going somewhere different. That signal reaches the listener before the first question lands.
Who Gets the Most from This Skill
Independent podcast hosts managing their own show production. Content creators who interview guests as part of a broader video or audio strategy. Founders and professionals who record thought-leadership episodes but don't have a dedicated production team to handle research and scripting.
The biggest pain for independent creators isn't the recording — it's the prep time. A well-researched episode can take three to four hours to prepare properly: reading prior interviews, pulling clips, building a question sequence that actually goes somewhere. That work is what separates the shows that grow from the shows that plateau. The skill compresses that research cycle without cutting corners on what it surfaces.
It's also useful for anyone who records episodically but inconsistently — where the quality of your script reflects how much energy you had that week rather than how good your process is. A skill-driven workflow makes the floor higher on a bad week.
The Output You Walk Away With
The skill delivers a complete episode script: a fully written intro block (not a fill-in-the-blanks template, but prose you can read or adapt), a guest bio that highlights the specific background relevant to this episode's angle rather than reciting their full career, a structured question sequence of eight to twelve questions with suggested follow-up hooks embedded, and a closing segment with a clear sign-off that connects back to the episode's central theme.
From there: paste it into your recording notes, read from it directly, or use it as a structured prompt for your own ad-lib style. The script is a scaffold, not a cage. Most creators end up using sixty to seventy percent of it verbatim and improvising around the edges — which is exactly how a good prep document should work. You know where the episode is going. The guest doesn't feel like they're being processed.
Every podcast listener has heard the episode that was clearly underprepared — where the host asks the guest to summarise their entire career in the first three minutes and the conversation never recovers its footing. The research is the preparation, and the preparation is what separates the interview that gets shared from the one that just gets archived. The writing is the last step, not the first.
The next piece most people tackle from here is show notes formatted for both listeners and search. If you're working across the full Video & Pod workflow, the Video & Pod bundle covers everything in one place.