# Create an MP3 voiceover for a product demo

> Write a short walkthrough script, turn it into an English MP3, then pair it with a screen recording. Includes real audio, footage and an editable example.

- Published: 2026-11-11
- Reading time: 4 min (~758 words)
- Tags: text to speech, screen recording, Studio
- Canonical: https://tautau.xyz/blog/create-mp3-voiceover-product-demo

## Key takeaways
- Write for the action on screen and leave room for the viewer to follow it.
- Preview the voices, generate the script and download the MP3.
- Import the recording and narration as separate tracks in Studio Preview.

A screen recording can show exactly where to click and still leave someone wondering why. A short voiceover gives the action a purpose. You can write it after recording, check the wording, then place the audio beside the footage.

This example pairs real tautau editor footage with an MP3 generated by the [text-to-speech tool](/tools/text-to-speech). It is a small demonstration you can download and try yourself.

## Start with the finished example

<video controls preload="none" poster="/videos/features/03-choose-a-clip-and-export.jpg" aria-label="Editor footage with an example text-to-speech narration">
  <source src="/videos/examples/product-demo-voiceover.mp4" type="video/mp4" />
  <track kind="captions" src="/videos/examples/product-demo-voiceover.vtt" srcLang="en" label="English" default />
  <a href="/videos/examples/product-demo-voiceover.mp4">Download the narrated example</a>
</video>

The picture shows the recording editor. The separate narration explains the action:

> Move to the part you want to keep. Mark the start and end, add the clip to the queue, then export it. The original stays intact.

Download the [original screen footage](/videos/features/03-choose-a-clip-and-export.mp4), the [Aria MP3 narration](/audio/tts/product-demo-voiceover.mp3), or the [finished reference video](/videos/examples/product-demo-voiceover.mp4). The reference combines those two files; you can make your own version in Studio Preview using the steps below.

## 1. Write for one action

Pick the moment the viewer needs to understand. For this clip, it is choosing a useful range and exporting it while keeping the original. A feature list would add more words without helping someone follow the screen.

Write the script in the order the action happens. Name the control when the reader needs to find it. Leave a beat after a significant step. If the recording moves too quickly for the explanation, shorten the script or give the picture more time.

For a longer walkthrough, use the [speaking-time calculator](/tools/speaking-time-calculator) to estimate the first draft. Enter your own pace or work backwards from the time available. The estimate excludes pauses, so the generated audio is the timing you should use for the edit.

## 2. Generate the MP3

Open [text to speech](/tools/text-to-speech) and paste the script. The prerecorded voice previews use the same sentence, making it easier to compare Aria, Atlas and Nova. Preview playback does not spend your character allowance.

Choose a voice and press **Play**. Listen for names, numbers and abbreviations. A written phrase such as “v two point three” may read more naturally than a compressed version label. Change punctuation or split a long sentence if the delivery feels rushed.

When the read sounds right, choose **Download MP3**. Replaying or downloading the same generated text and voice in the current tab reuses that audio. Changing the text or voice requires another generation.

Each conversion accepts up to 4,000 characters. The anonymous trial has 5,000 characters in total; a free account includes 50,000 characters per month. Those are character allowances, not a promise about how many minutes every script will produce.

## 3. Put the picture and sound on separate tracks

Open [Studio Preview](/studio). Choose **Import media**, or drop the video and MP3 over the editor. Each valid source becomes an editable track. Keep both original files beside your project backup.

Turn off **Sync matching audio** for this import. A newly written voiceover is different audio from the recording, so there is no shared spoken event for automatic matching to find. Place it manually instead.

Select the narration clip and adjust **Timeline start** in the Inspector. Move it until the words and the action make sense together. If the screen recording already contains microphone audio, turn off **Include clip audio** for the unwanted source or mute its track. Listen to the actual mix rather than relying on the shape of a waveform.

Studio remains in Preview. The [Studio guide](/docs/studio) documents supported formats, local project storage and current limits. The [recording-editor demonstration](/demos/edit-a-screen-recording) shows the separate workflow for selecting and exporting clips from a captured recording.

## 4. Watch once without touching the controls

Play from before the first action to after the final sentence. Check whether the viewer has enough time to find each control. A pause in the narration can be useful; a sentence describing a screen that has already disappeared is harder to follow.

Check the opening and ending too. The explanation should start after the picture establishes the task and finish before the video cuts away. Turn the volume down for one pass and make sure the written labels or captions still communicate the action.

For the Studio version, choose **Video · WebM** to export the picture and mixed audio. WebM export runs in real time, so keep the tab visible until it finishes. The downloaded MP3 remains useful on its own, and can also go into another editor if you need a different final container.

Use **Save project** for an edit backup. Keep the source video and narration together so you can revise a sentence when the interface changes, without starting the recording again.

---
Markdown twin of https://tautau.xyz/blog/create-mp3-voiceover-product-demo — provided for LLM/agent consumption. tautau is a speech-to-text Chrome extension and web app: https://tautau.xyz
