Script & voice
Write, edit and regenerate the narration script, use what you said while recording, and choose the AI voice and tone for a video.
The script is what the AI voice says in your video: a line for each step, spoken while that step is on screen, plus an opening and a closing line and short lines for what happens between steps. You can let the AI write it, edit any line by hand, or have it cleaned up from what you said while recording. The Voice tab sets who reads it and in what tone.
Write the first script
The script is written after you've checked the steps, so each line describes the right click.
- Open the video in the editor. It opens on the Hotspots tab.
- Select any step and click Write voiceover script at the bottom of the panel.
The footer shows Writing voiceover script…. When it's done, the editor switches to the Script tab. Until then, that tab says "No script yet".
The AI looks at frames of your recording and writes:
- A line for each step, one or two full sentences sized to how long that step's screen is visible: what you do there and, when the screen shows it, what it's for.
- A short line after a step where something worth saying happens before the next one, such as an AI drafting a page or a form being filled in. It's spoken over that stretch while the Studio style speeds it up.
- An opening line spoken over the title card, and a closing line spoken over the end card. They're only heard when the video has those cards.
It only describes what's on screen and what your context says, and uses the button and field names as they appear. The Demo kind and Context you gave when creating the video shape the wording. See Create a narrated video.
Edit a line
On the Script tab, each line has its own box. A step's line is headed with the step number, the step title and the time it starts (for example "@12.4s"). The other lines have a dashed border and say where they're spoken: Opening line · over the title card, After step 3 · while the next step loads and Closing line · over the end card. Clear a line's text to leave it out of the video.
- Click into a line and change the text.
- Click the play button next to the time (Hear this line) to hear it in the video's current voice.
Edits save automatically; the top bar shows Saving… and then Saved. They reach the finished video on the next render.
Keep a step's line about as long as its screen is up. When a line takes longer to say, the video freezes on that screen until the line finishes; when it's much shorter, the rest of the screen time plays without a voice (sped up, with Tight pacing). See Steps, zooms & timing.
Regenerate the script
Click Regenerate script at the bottom of the Script tab to have the AI write every line again from the current steps. Do this after you add, remove or retitle steps, or after retiming a step, since each line is sized to its step's screen time.
Regenerating replaces every line, including lines you edited by hand. Copy any wording you want to keep first.
Use what you said while recording
If you talked through the steps while recording, TurtleScene can turn that into the script. It transcribes what you said, matches each part to the step that was on screen, and cleans it up: no "ums", no repeats. The AI voice still reads it; your own voice never appears in the finished video.
- Record with Record my microphone switched on in the recorder, and talk through the flow as you go.
- In the editor, open the Script tab.
- Click Use what I said while recording.
The footer shows Listening to your narration…, then the lines appear: one per step, without the opening, closing and in-between lines the AI adds. Edit them like any other line.
If the recording has too little speech, you'll see "We couldn't hear any narration in this recording." Record again with your microphone on, or click Regenerate script to let the AI write it.
This also replaces every line in the current script.
Choose a voice and tone for this video
- In the left rail, click Voice.
- Under Voice, click the play button next to a voice to preview it, then click the voice's name to choose it.
- Under Tone, choose how it should sound.
The choice saves straight away ("Saved. Applies to the next render.").
Voices: Alloy, Ash, Ballad, Coral, Echo, Fable, Nova, Onyx, Sage, Shimmer, Verse, Marin and Cedar.
| Tone | Sounds like |
|---|---|
| Friendly | Warm walkthrough, not salesy |
| Professional | Calm, credible, enterprise |
| Energetic | Upbeat launch video |
| Calm | Patient tutorial |
| Authoritative | Keynote, bold claims |
Workspace defaults and per-video settings
Every video starts with Workspace default selected for both voice and tone, so it uses the narrator set for the whole workspace. Choosing a voice or tone on the Voice tab overrides the default for this video only. Choose Workspace default (voice) or Default (tone) again to go back.
To change the default for every video that hasn't overridden it, go to Settings in the sidebar and use the Voice & narration card. See Voices & pronunciation.
Fix how product names are said
If the voice mispronounces your product name or a term, add it to the Pronunciation dictionary in Settings › Voice & narration: write the term as it's spelled and how it should be said. It applies to every video and story in the workspace, so you don't need to misspell words in your script. The workspace settings link on the Voice tab takes you there.
The pronunciation dictionary is available on Starter and up.
Re-render after changes
Script, voice and tone changes reach the video on the next render. Click Re-render in the top bar, or Generate voiceover & render MP4 at the bottom of the Script tab. Lines you haven't changed reuse their audio, so re-rendering after small edits is quick. See Render, download & share.