Voice Over Sample Scripts by Genre: Original Examples

Direct answer: a voice over sample script is a short, ready-to-adapt piece of spoken copy written for one job: selling, explaining, narrating, instructing, or announcing. The right sample depends on the genre, because a 20-second ad, a chapter of an audiobook, and a phone menu prompt follow different rules for length, pace, and tone. This guide gives you one original sample script per genre, a comparison table to pick the right shape, and a checklist to adapt any of them.

Every script below is original and written for this guide. The businesses, products, people, and lines are fictional, and none of them describes a tested real product or a measured customer result. Treat them as shapes to copy, not as copy to run unchanged. The word-per-minute figures in the comparison table are planning estimates for budgeting time, not measured performance values. Time the finished audio before you rely on any duration.

What makes a voice over sample script work

A usable sample does five things at once. It names one audience, gives that audience one job, sets a length that fits that job, holds a single tone, and lands on one action. A script that tries to sell, explain, and reassure in the same breath usually does none of them well. Genre is simply the first filter that tells you what those five choices should look like.

Two production habits carry across every genre. First, keep the spoken line clean and move performance directions (such as slower or warmly) into a separate file, because a human reads every cue while a machine reads only the words and the tags its model supports. Second, write short sentences and read them aloud early. A line that looks fine on screen often shows a stumble or a hiss of breath that only appears when spoken.

• One job per script, stated in your own planning notes before you write a word.
• A length target taken from the genre, not from how much you have to say.
• A single tone, chosen once, then held for the whole read.
• One call to action or one next step, never two competing ones.
• A clean spoken layer, with any performance cue kept out of the words a machine will read.

Seven genres, seven original sample scripts

Commercial and radio spots

A commercial has to be understood on one listen, and often while the listener is doing something else. Lead with the offer, keep one brand name, and place the action at the end so it is the last thing heard.

The walls are still bare, and the paint is still wet. Nearline Hardware has brushes, rollers, and trays stacked by the door. Bring your color card, get the match, and leave with the tin. Nearline Hardware on Fifth Street. Open until eight.

Original fictional commercial script, written for this guide

Notice that the sample never claims a price, a discount, or a result. It gives the listener something to picture, names the place, and states the hours. When you write your own, keep the same discipline: describe what is true today rather than promising an outcome. For a full set of audio-only structures, the radio ad script examples show how a spoken-only brief can carry a message without a screen.

Explainer and product demo

An explainer works from a problem to a visible result. Name the job, show the shortest path through it, then stop. Keep the words and the screen in step, because a claim the viewer cannot see is still just a claim.

If your team keeps losing track of who approved what, this is the fix. Vellum Board puts every request on one page. Open a card, tag the reviewer, and the status changes as they move it. No separate email thread, and no message asking for an update. See the public tour to try one request yourself.

Original fictional product demo narration, written for this guide

Pair each spoken line with something the viewer can check. If the voice says the status changes, the capture needs to show it changing. The product demo script template carries the same structure as a beat sheet, which is easier to hand to an editor than a wall of prose.

Audiobook and long-form narration

Long-form narration is a marathon, not a sprint. The voice should disappear so the story stays in front. Favor steady volume and steady pace over dramatic swings, and keep names and recurring words pronounced the same way chapter after chapter.

The house on Marrow Lane had one rule, and my grandmother repeated it every summer. Never answer the door after the streetlights come on. I broke that rule the year I turned twelve, and I have spent a long time deciding whether I regret it.

Original fictional audiobook passage, written for this guide

Audiobook projects also have platform rules that sit outside the script. For example, ACX's published audio submission requirements call for consistent loudness (their page lists a range of -23 dB to -18 dB RMS), peaks under -3 dB, a noise floor no higher than -60 dB RMS, and 192 kbps or higher MP3 at 44.1 kHz, so plan the read with those technical targets in mind. ACX's help pages also state that standard submissions must be human-narrated unless otherwise authorized, so check the current requirements before you plan an AI narration for that channel. The audiobook narration workflow covers the chapter-by-chapter pass that keeps a long book consistent.

E-learning and course narration

Course narration teaches. It is calm, it repeats key terms, and it never races ahead of the screen. Write one idea per screen and tell the learner what to watch for before you explain it.

In this module you will measure a rising water table three ways. Watch the diagram as each method appears, and notice where the reading is taken. When you are ready, use the check on the right to test one measurement. If it is low, replay the second step before you move on.

Original fictional course narration, written for this guide

The repetition is deliberate: a learner who is reading the screen may miss the first pass. Keep each line short enough to finish before the visual changes. The course narration workflow covers exporting one audio file per lesson so a single changed screen does not force a full re-record.

Phone prompts, IVR, and voicemail

A phone prompt is heard by someone who is already mildly annoyed. Say the action before the key, keep greetings short, and never stack more than a few options in one menu.

Thank you for calling Northgate Clinic. For appointments, press one. For test results, press two. To reach the front desk, stay on the line. Please have your patient number ready before you are connected.

Original fictional IVR script, written for this guide

Order matters here: the listener needs the destination before the key press, not after. If your system can only play pre-rendered clips, write each prompt as its own short file so a wording change does not force you to re-render the whole menu. The voicemail and greeting script examples cover the shorter one-step versions of the same pattern.

Podcast intros and show opens

A podcast intro does three jobs in under a minute: it names the show, states the promise, and sets up the episode. Energy should come from the words, not from rushing them.

This is Field Notes, the show where working scientists explain the story behind a single number. Today, a count that nobody could explain for forty years, and the clerk who finally did. I am your host. Let us get to it.

Original fictional podcast intro, written for this guide

Keep a fixed set of intro lines for the show and swap only the episode-specific sentence. That makes a repeatable template and a consistent brand sound across weeks. More variants, including interview and solo formats, sit in the podcast intro script examples.

Game and character lines

Character lines are short, repeatable, and delivered in a consistent voice. Write to the character's point of view, give the player a clear direction, and expect each line to be heard many times, so it must survive repetition.

Ho there, traveler. The bridge to the east washed out in the spring floods, and the ferryman will not cross after dark. Take the old road instead. Follow the lanterns, and do not stray from them. If you find the mill, you have gone too far.

Original fictional character script, written for this guide

Loops are the test. A line that sounds fresh on the first play can turn grating by the tenth. Keep each line under two sentences, and record or generate a small set of alternate deliveries so the same sentence is not repeated word for word every time.

Genre comparison table: pick the right shape

Use the table to set a target before you write. The narration column is a planning estimate only, based on a comfortable reading pace, and it is not a measured performance number. Actual timing depends on the voice, the pauses, and the edit.

How to adapt a sample script to your own project

A sample is a starting shape, not finished copy. Run it through the same short pass every time so the result fits your brand and your genre.

1. Name the one job and the one audience in a single sentence you keep at the top of the file.
2. Set a length target from the genre table, then cut the smallest detail that does not serve the job.
3. Write the spoken line clean, and move every performance cue into a separate notes column.
4. Read it aloud once at your intended pace, and mark any word that catches or sounds unnatural.
5. Write down pronunciations for names, numbers, and acronyms so a second take sounds the same as the first.
6. Split the script by speaker if more than one voice speaks, and assign one voice to each block.
7. Generate or record a short test of each voice, then confirm the voices are distinct before the full pass.
8. Time the finished audio against your target, and cut a repeated idea before you speed up the delivery.

Keep a copy of the original sample next to your finished script. When a client asks why a line changed, the comparison shows exactly which genre rule drove the edit, which is faster than rebuilding the reasoning from scratch.

Where a local voice studio fits, and where it does not

Once the words are settled, Murmur can turn a clean script into audio on an Apple Silicon Mac running macOS 15 or later. It is a local-first voice studio for text to speech, voice cloning, voice design, multi-speaker projects, and export, and it runs the models on your own machine. That means no per-character fees and no upload of your script to a cloud service, which helps when a sample is under review or a client asks for a private read. Murmur is $49 one-time, has a free plan and a 3-day Pro trial, and includes a 7-day refund policy.

Genre shapes the model choice. For a plain commercial or course read, a bundled preset voice is usually enough. For a two-hander, an audiobook cast, or game dialogue, the multi-speaker workflow lets you assign a different voice to each block and render each part as its own track so one line can be retimed without redoing the scene. Some local models in Murmur accept a small set of inline reaction tags typed straight into the script, such as a laugh or a sigh, while other variants hide those tags, so keep the performance cues in brackets on your master and check the model list in the app before you rely on a tag. Because model support varies, remove any tag a model does not understand so it is not spoken as a word. Voice cloning needs the permission of the person whose voice you use, and a disclosure note is worth planning before any public release.

Murmur is not the right tool for every genre. It is a Mac app, so if you work on Windows, want live or real-time audio for a voice agent, or need several people editing one shared cloud project, a cloud service fits that workflow better. Murmur also does not do speech to text, so it cannot turn a scratch recording into a script, and it does not stream lines in real time for a running game. It is strongest when you already have the words and want private, repeatable audio you can regenerate from the same script.

Frequently asked questions