Why use text-to-speech for a video
For an app demo, a tutorial or a short ad, the voice has one job: say clearly what is on screen. A synthetic voice does that consistently, you can change a word without re-recording the whole take, and you can produce the same video in another language by swapping the script. The trade-off is warmth: a person talking to camera feels more personal, which is why many ads open with a real face and hand the explanation to a voiceover.
Option 1: the say command in Terminal
macOS ships with a speech synthesizer you can call from Terminal. This writes a sentence to an audio file:
say -o line1.aiff "Track every habit in one tap."
Add -v with a voice name to pick a voice (say -v '?' lists the installed ones). Then import the files into your video editor and line them up by hand. It is free and private, but every sentence is a separate file and the captions are up to you.
Option 2: a cloud text-to-speech service
Online services offer many voices and styles. You paste the script, download the audio and place it in your editor. They usually charge by character or by month, and your text is sent to their servers, which matters if the script describes an unreleased product.
Option 3: text-to-speech inside the editor
When the editor generates the voice itself, the script lives on the timeline: change a line and the audio and captions follow. This is how ClipForge works on the Mac.
- Write one short line per scene: the hook, each proof shot, the call to action.
- ClipForge synthesizes each line on your Mac and places it on an audio lane under its scene.
- Captions are timed from the speech engine's own word timings, so each word lights up as it is spoken.
- Music ducks under the voice automatically.
- Export one project to TikTok, Reels, Shorts, YouTube and feed formats.
Nothing is recorded and nothing is uploaded; the voiceover is part of the free editor.
How to write a script a synthetic voice reads well
- Short sentences. One idea per line; a period is a natural pause.
- Write numbers the way you want them said. "Ten minutes a day" rather than "10 min/day".
- Spell out abbreviations the voice might read letter by letter, or say them as words.
- Read it aloud yourself first. If you stumble, the voice will sound stiff there too.
- Leave room for the visuals. A three-second shot fits about eight words.
When to record your own voice instead
If the video is about you, a personal story or a founder update, record it. A real voice with small imperfections builds trust in a way a synthetic one does not. For product demos, explainers and ad variations you plan to test, text-to-speech is faster to change and easier to keep consistent.
Questions
Can I add text-to-speech to a video on a Mac for free?
Yes. macOS includes system voices and the say command can save speech to an audio file you then place in any editor. ClipForge's on-device text-to-speech is part of its free editor.
Does on-device text-to-speech upload my script?
No. On-device means the voice is generated on your Mac. Cloud text-to-speech services receive your text on their servers to synthesize it.
How do I sync captions to a text-to-speech voiceover?
Either transcribe the generated audio afterwards, or use a tool that times the captions from the speech engine itself. ClipForge does the second: captions follow the engine's own word timings, so they match the voice exactly.
Which voices and languages are available?
On a Mac that depends on the voices installed in System Settings (Accessibility › Spoken Content). Download more voices there, including higher-quality versions, and they become available to apps that use the system voices.