Plenty of people do not want to record their own voice. They are not sure about their accent, they have no quiet room, or they just dislike the sound of themselves. A generated voice solves that in a minute. It also has costs. Viewers notice it, and for some videos it is the wrong choice. This guide covers how to make a voiceover from a script in Superframes, which videos suit a generated voice, which ones need you, and how to get the best from either.
A generated voice does well when the viewer cares about the information and not about who is giving it. Explainers, screen walkthroughs, slide decks, how-to steps and short announcements all fit. It also helps when you need to change the script often, because fixing a sentence means generating one line again, not setting up a microphone. It helps if you need a second language you do not speak. Read the result with someone who does before you publish, though. A voice can read a word wrong and you will not hear it.
Record your own voice when the video is about you or asks for trust. A founder note, a personal story, a review, a message to customers, a tutorial where you are the teacher: viewers want a person, and a generated voice reads as a stand-in. It also loses the small things that make a real voice, like a breath before a hard point or a laugh. If you doubt your voice, remember that a real voice with a few stumbles often beats a flawless generated one. The editor can help you with that: cut the silences, tighten the gaps and clean the noise, all on your device.
Open the Generate window and pick the Voice tab. Paste your script, choose a voice, and check the price before you press Generate. The audio comes back to your device and you place it on an audio lane under your video, then line it up with what is on screen. For a long script, generate it in parts, one section at a time. A part that comes out wrong then costs you a small amount to redo, and it is easier to match each part to a scene. Because the price depends on the length of the text, cutting a rambling script is also the cheapest way to improve the result.
A script that reads well on a page often sounds stiff aloud. Use short sentences. Use contractions. Spell out numbers and abbreviations the way you would say them, such as "twenty twenty six" if that is how it should sound. Put a comma or a full stop where you want a pause. Then read it aloud yourself, once, at the pace you want. Wherever you run out of breath or trip, the voice will sound odd too. Fix the sentence, not the voice.
Once the audio is on the timeline, move your clips so the picture matches the words. When the voice says "now open settings", the settings screen should already be there. Cut the pauses that feel too long and let the ones that give a point weight stay. Add captions too. Superframes transcribes on your device, so the captions come from the audio without the file travelling anywhere, and many viewers watch with the sound off.
Some platforms ask you to label content made with AI, and viewers tend to trust you more when you are open about it. Check the current rules of the platform you publish on, because they change. Never use a generated voice to pass as a real person who did not say the words.
Voice is 9 credits for every 1,000 characters, and 100 credits is one US dollar. The price shows before anything is made, and a failed or cancelled generation costs 0. When you generate, the words to speak go to our AI providers. Your video files stay on your device.
Add auto-captions to any video in your browser. On-device transcription with viral caption styles, no upload, your footage and audio stay private.
Clean up muddy audio and reduce background noise in a video, in your browser. No upload, your video and audio files stay on your device.
Swap a video’s audio track with music or voiceover from another clip, in your browser. Video stays untouched and is never uploaded.
Yes. Open the Generate window, choose Voice, paste your script and press Generate. The audio comes back to your device and you place it on the timeline.
Voice is 9 credits for every 1,000 characters of script. The price shows before it is made, and a failed generation costs nothing.
It depends on the video. For explainers and walkthroughs a generated voice is fine. For anything personal, or anything that needs trust, record yourself.
The words to speak. Your video files stay on your device, and nothing is sent until you press Generate.
It is a good habit, and some platforms require a label. Check the rules of the platform you publish on.