Podcast Script Example: A Complete Episode, Start to Finish

Short answer: A podcast script example is the full written plan for one episode: the exact spoken words plus production notes, divided into segments so each part has a job. What follows is a complete worked episode script for a 20-minute interview show, from the cold open to the outro, then a word budget to size each segment and a pre-publish checklist. Copy the structure and adapt the details rather than the other way around.

The show, host, guest, company, and any quotes or figures below are fictional and original. They exist to demonstrate script mechanics, not to report a real episode. If you want the reusable header and the beat fields first, the voice over script template covers that worksheet, and the podcast intro script examples go deeper on the opening alone. This page is the whole episode in order.

How a scripted episode is divided

Treat the table as a skeleton, not a cage. A solo explainer might drop the interview segment and lengthen the takeaway. A narrative show might replace every label with scene beats. The point is that each part earns its minutes, so you never record a segment that has no job. Apple's guidance to present your podcast recommends writing to match your show's vocabulary, energy, and personality, and to lead an episode title with its most eye-catching information. Apply the same idea to every spoken line: the words should sound like this show, not like a generic announcer.

The complete worked episode script

Here is the whole script for episode 12 of an invented show called Ground Truth, about how small teams make decisions with data. Host Priya Raman interviews guest Marcus Bell about a reporting project that went sideways. Speaker labels and bracketed production cues guide the edit; only the spoken text is recorded. Read it top to bottom as the finished episode.

Cold open

PRIYA (voice over a clip): The sales team had already moved on. The dashboard still showed the old forecast. MARCUS (clip): We were not arguing about the numbers. We were arguing about which numbers anyone believed. [COLD OPEN MUSIC IN, THEN UNDER]

Original fictional cold open

The cold open spends thirty seconds on a specific moment instead of a summary. It pairs one host line with one guest clip, so the listener hears the tension before the show even names itself. Keep the clip short enough that the promised explanation still feels new. If you cannot find a real moment, this is a sign the episode may need more reporting before it needs a script.

Intro and housekeeping

[INTRO MUSIC UNDER, THEN OUT] PRIYA: You are listening to Ground Truth, a show about the small decisions behind working with data. I am Priya Raman. PRIYA: Today, why a team kept trusting a spreadsheet over the dashboard they had just paid to build, and what finally changed. Our guest ran that project from the inside. PRIYA: Quick housekeeping before we start. The transcript and show notes are linked on the episode page. Now, the room where it went wrong.

Original fictional episode intro

Notice the order: show name, host name, a specific reason to listen, one piece of housekeeping, then straight into the content. There is no "welcome back to another amazing episode." The promise "why a team kept trusting a spreadsheet" is a real question, so the listener knows what the next twenty minutes answer. Move anything that is not that promise, such as sponsor reads or follow requests, to a place where it does not delay the first useful moment.

Segment one: set the scene

PRIYA: Marcus, set the scene. It is March, the dashboard is live, and the sales team will not open it. What did the room actually look like? MARCUS: Picture a Monday forecast call. Nine people, one shared screen, and a number two people had already disagreed about before the call began. PRIYA: So the tool was not the problem. MARCUS: The tool worked fine. The problem was that nobody could say where the number came from. When you cannot point at the source, you go back to the file you built yourself.

Original fictional interview segment

The host asks one clear question and then gets out of the way. The guest answers with a picture, not an abstraction: a Monday call, nine people, one contested number. A strong scene-setting answer gives the listener something to hold. If your draft of this segment is a paragraph of explanation, split it into questions the guest can actually answer in their own words.

Segment two: the turn

PRIYA: What did you try first, and why did it not stick? MARCUS: We added more charts. That made it worse. More views meant more versions of the truth on more screens. PRIYA: What actually changed it? MARCUS: We stopped shipping a number and started shipping the definition. One line at the top of each card said what counted and what did not. Less impressive, far more trusted. PRIYA: Give me the smallest example. MARCUS: A closed won deal is anything with a signed order in the system, not a verbal yes on a call. Writing that down ended an argument that had run for months. PRIYA: And the lesson is not "build better dashboards." MARCUS: The lesson is that trust attaches to a definition, not a chart. If two people define a metric differently, the prettiest chart loses to that disagreement every time.

Original fictional interview segment

This is the longest segment because it carries the episode's turn: the failed fix, then the change that worked. Note how the host keeps asking shorter and shorter questions to pull out specifics. The example about a signed order versus a verbal yes is concrete enough to be repeatable, which is what an interview clip needs to survive outside the episode. Avoid inventing quotes or outcomes for a real guest; if the material is thin, ask again rather than write it for them.

Segment three: the takeaway

PRIYA: If a listener runs one of these projects next month, what is the one thing to do first? MARCUS: Write the definition, put it where the number lives, and make one person own it. Do that before you pick a chart type. PRIYA: So the script for the listener is smaller than the script for the software. MARCUS: Much smaller. One sentence, in the right place, with a name next to it.

Original fictional closing segment

The takeaway segment converts an interesting conversation into a single action. It repeats the episode's core idea in plain language and gives the listener something they can do on Monday. Resist the urge to add five more tips here. If the interview produced several good ideas, keep the strongest one and let the others become a separate episode.

Outro and credits

[OUTRO MUSIC UNDER] PRIYA: That is episode twelve. Marcus Bell leads reporting at a company we invented for this example, so treat his method as a prompt, not a testimonial. PRIYA: If this changed how you read your own dashboard, send the episode to the person who argues with you about it. Next week, what happens when the definition is right and the data still arrives late. PRIYA: Ground Truth is produced and edited in house. Music by our fictional composer. Thanks for listening. [OUTRO MUSIC UP AND OUT]

Original fictional episode outro

The outro does four things in under a minute: it labels the episode number, credits the guest, teases the next episode, and closes the credits. The line that separates a method from a testimonial matters. A script is where you control that framing, because once the audio is published you cannot walk the claim back.

Word budget for a 20-minute episode

A word budget is the fastest way to keep an episode from drifting long. The table below assigns each segment a target range at an assumed 150 words per minute, a common planning figure for clear conversational speech. To estimate a duration at that rate, divide the spoken words by 150 and multiply by 60. These are planning figures, not measured performances, and they do not include pausing, laughter, music, or turn changes.

Set the budget before you record and check it after the edit. If the episode runs long, cut a whole segment rather than raising the host's speaking speed, because a rushed host is harder to follow than a shorter show. If a segment comes in far under budget, ask whether it still earns its place. A planning estimate is not a contract; it is a way to notice drift early.

Turn the draft into publishable audio

1. Record or generate one segment at a time and keep clean per-speaker files so a single line can be replaced later.
2. Listen once only for pronunciation of names, brands, and numbers before you fix anything else.
3. Regenerate or re-record only the lines that are wrong instead of rebuilding the whole episode.
4. Assemble timing, music, and levels in your editor so the written script stays the source of truth.
5. Export a working master, then a separate delivery file in the format your host expects.

If you generate narration on a Mac, Murmur can render each speaker as a separate track from the same script and re-render one line without touching the rest. It runs locally on Apple Silicon Macs running macOS 15 or later, so scripts and cloned voice samples stay on the machine, and it is a voice tool rather than a full audio editor. That focus is also its limit: Murmur will not host, mix, or publish your show, and it is not the right choice if you need Windows, a phone app, or a live streaming voice. The multi voice podcast workflow covers the speaker-by-speaker version of this process.

When the episode is assembled, export using a format your host accepts. The voiceover export guide explains the difference between an editing master and a delivery copy. Apple lists the audio and video formats it accepts for feed episodes on its create an episode page, and it separately requires an episode number for serial shows but treats it as optional for episodic ones.

What to check before you publish

• Confirm every name, title, and company against the guest's own spelling, not your draft.
• Cut any claim you cannot show, including invented listener numbers or invented review quotes.
• Read the intro and the outro out loud; if you stumble, the host will too.
• Re-check the definition or technical line in any segment, because that is where a small error does the most damage.

Publish a transcript alongside the audio. W3C's guidance on transcripts for audio and video media states that prerecorded audio-only content, which includes podcasts, needs a transcript to meet the lowest level of the accessibility standard, and that a transcript can also add headings and speaker labels to make the episode easier to scan. Generating the transcript from the finished audio, rather than from your draft, keeps it honest about what was actually said.

Frequently asked questions