Turn a screen recording into a walkthrough your customers can follow. A clear voiceover tells them what to do, where to look and why the result matters, without making them guess what each click means.
If you are creating a product demo or onboarding video, ElevenLabs is worth considering when you want to work from a script rather than record your own narration. Its Studio editor lets you generate speech, arrange it alongside your recording and revise individual lines. That is useful when a button label changes or an explanation needs tightening. You still need to check the voice and match the timing to the screen.
ApexTool may earn a commission from purchases made through links in this guide.
What you need
- A screen recording that shows one task in your software.
- A short script describing the actions in that recording.
- An ElevenLabs account with access to Studio.
Use a sample workspace without customer information. If you have not recorded yet, pause briefly before each action and after its result. That gives you room to fit the narration later.
For a business demo, generate the final audio on a paid plan. Free-plan audio has no commercial license, and upgrading later does not change that. Paid-plan rights exclude Beta Services and depend on having the necessary rights to your content. Check the publishing rules and current pricing before generating production audio. New takes can use credits, so check the displayed charge.
1. Write for the recording you actually have
Watch your recording and note where each action starts. Write one short line for each action, using the labels visible in your software.
Here is a sample script for a report-export demo. Adapt the words and actions to your own product:
| What the viewer sees | What the voiceover says |
|---|---|
| The report, with Export visible | “Need this report in a spreadsheet? Start with Export.” |
| The export menu opens | “Choose CSV, then select Download.” |
| The downloaded report opens in a spreadsheet | “Your report is ready to sort, filter and share.” |
On a small screen, scroll the table sideways to read every column.
The opening gives the viewer a reason to watch. The middle explains the action. The final scene shows the payoff. Keep scene notes out of the spoken script, and only describe results your software actually delivers.
Do not assign a fixed duration yet. The finished voiceover and the time needed to follow the screen will determine the length.
2. Create the video project and your first voiceover
In ElevenLabs, open Studio:
- Choose + New blank project → Video project (New).
- Use Library → Create + → Upload to add your recording, then drag it onto the timeline.
- Choose Create + → Speech, enter one line and select a voice.
- Generate the speech clip and drag it onto a separate track alongside your video.
Start with a line containing your product name or a difficult acronym. Listen before generating everything else. Choose a voice that suits the explanation, not just one that sounds impressive in a preview. If the delivery feels rushed, shorten the sentence first.
Use Script → Voiceover to edit and regenerate narration. For a recurring pronunciation problem, add an alias under Script → Pronunciations. See the Studio instructions for the controls.
3. Match the narration to the screen
Play your first scene. Move its narration clip until the instruction accompanies the relevant action. The control should already be visible when it is named, and the result should stay visible after the click.
In the sample, “Choose CSV” belongs with the export menu. The final line belongs with the spreadsheet, not a loading screen.
A timing fix you can actually use
Suppose an instruction lasts eight seconds, but the recording leaves that screen after six seconds. Moving the audio earlier may simply create a different mismatch.
Try this order:
- Remove unnecessary words and generate the shorter instruction again.
- If it still needs more time, use more of the pause you recorded after the action.
- If there is no usable pause, re-record that small section and hold the result longer.
Once the first scene works, add the next clip. After changing a line, replay its transition into the following scene too. A new take can be longer or shorter than the one it replaces.
If the whole video feels rushed, reduce the number of actions. Faster speech will not make a crowded screen easier to understand.
4. Add captions and export
Open Captions, select the narration track as the source, then correct the transcript and adjust the caption style. Select Export when ready.
Watch the downloaded file once with sound and once muted. Check that:
- Instructions match the action, including the last step.
- Captions spell product names and button labels correctly.
- Captions do not cover the controls being demonstrated.
- The result stays visible long enough to understand.
- No private information appears in the recording.
Keep the final script with the project so you can replace individual lines when the software changes.
Try it with one scene
ElevenLabs makes sense for a scripted walkthrough you expect to revise. If the personal connection with a founder or trainer is the point of the video, recording your own voice may be the better choice.
Start with the report-export example or one scene from your own demo. Listen for clear product names and check that each instruction lands with the right action before building the rest.