Sound

This guide covers music (composed with AI, from the VibeDay music library, or your own), voice-overs (typed or recorded) and sound effects (made with AI or your own), cleaning up a voice, and removing ums and long pauses.

Open the Audio tool on the left. It has three tabs: Music, Voice-over and Effects.

Compose a music track with AI

Use this for an original instrumental track made to fit your video.

1. Click Audio and stay on Music.
2. Under Compose a track with AI, describe the mood, or tap a starter: Warm acoustic, upbeat, Calm lo-fi beat, Energetic pop, driving drums, Cinematic and hopeful or Playful and quirky.
3. Check Length. It starts at your video's length.
4. Click Compose. The price shows on the button before you start, and you're only charged when the track is ready.

The track appears straight away on a Music track at the start of the video, striped and labelled ⟳ Composing your track…. It usually takes under a minute. You can keep editing, and it plays as soon as it's done.

Good to know:

• Tracks are instrumental, composed just for your video.
• Music is lowered automatically while someone is speaking.
• If a track can't be made, it's taken off the timeline, nothing is charged, and a note says why. Try describing the mood a little differently.

Listen to your takes, or make another

Every track composed in this project is listed under Recently composed, newest first, with ▶ and its length.

• Press ▶ to listen. Nothing changes on the timeline until you choose.
• Click Use to make that take the project's music. It takes the place of the current music track where it starts, keeping its volume and fades (if there's no music yet, it goes at the start). ⌘Z brings the old track back.
• Click Another take in this style to compose a fresh take from the same description, at the same length and price, without retyping. It's listed as take 2, take 3 and so on, with ⟳ Composing your track… while it's being made. Listen, then Use the one you like.

Every take is kept in your Library, whether you use it or not.

Music library

The music library has about 200 original instrumental tracks made for VibeDay with Google's music AI. They're sorted into the same 16 styles as the reel music menu, in four groups: Energetic, Calm, Cinematic and Fun & genre. Each style has short tracks (about 30 seconds) and long ones (about 90 seconds), so most videos fit without looping.

They're included, no credits: using a library track never costs anything.

1. Click Audio and stay on Music.
2. At the bottom, under Music library, click Browse the music library.
3. Find a track:
   • Type a mood or a style in the search box, like sunny, focus, cozy or piano. Every word you type has to match.
   • Or tap a group (All, Energetic, Calm, Cinematic, Fun & genre), then a style. The heading shows the style and how many tracks it has, like Acoustic · 13 tracks.
4. Press ▶ to listen. One track plays at a time; press it again to stop.
5. Click Use.

Use puts the track on the Music track, fitted to your video:

• If the track is longer than your video, it's trimmed to end with the video, with a gentle fade-out (about 2 seconds).
• If it's shorter, it plays again from the start, with a short fade where it joins, until the video ends. The last repeat fades out gently; if only a moment would be left, the music ends on the track's own ending just before the video does.
• It takes the place of the music already there (all of it, if that was a library track playing more than once), keeping its start, volume and fades. ⌘Z brings the old music back.
• Like all music, it's lowered automatically while someone is speaking.
• Library tracks play at an even loudness, so switching tracks doesn't jump the volume.

Good to know:

• Library tracks are used inside your videos. They aren't offered as downloads, and they don't appear in your Library.
• The same tracks are available to every VibeDay customer; they aren't exclusive to you.
• Reels can use them too: on a finished reel, Change music has a Music library option. See the Reels guide.

Make a sound effect with AI

Use this for a short sound you don't have: a whoosh between scenes, a till opening, coffee being poured.

1. Click Audio, then Effects.
2. Under Describe a sound, tap a starter (Whoosh, Click, Pop, Applause, Cash register, Coffee pour, Notification ding) or describe it in your own words, like espresso machine steaming milk, close up.
3. Pick a Length: Auto lets the AI choose a natural length, or choose 3 s or 6 s.
4. Click Make the sound. The price shows on the button.

The sound appears straight away in the list under the button, labelled ⟳ Making your sound…. It usually takes a few seconds, and you can keep editing. When it's ready, press ▶ to listen, then Add to put it on an Effects track at the playhead. It plays as it is; drag its edges on the timeline to trim it. Every sound is saved to your Library as Sound effect · .

Price: 1 credit for every 3 seconds started: 3 s is 1 credit, 6 s is 2. With Auto you pay for the length that's made, at most 2 credits (the button says up to 2 credits). You're only charged when the sound is ready.

Good to know:

• Describe the sound itself rather than a brand, an artist or a song. If a description names one, we make the generic sound it stands for (a phone's ringtone rather than a particular brand's).
• If a sound can't be made, the list says why and nothing is charged. Try again, or describe it a little differently.

Add your own music or effects

1. Click Audio and choose Music or Effects.
2. Click a sound under Your music or Your sounds. These are the sounds in your Library.

Music goes after any music already there. Effects go at the playhead and play as they are. Each kind has its own named track.

Type a voice-over

Use this when you'd rather not record your own voice: type the words and an AI voice speaks them. These are the same voices our reels use.

1. Move the playhead to where the voice-over should start.
2. Click Audio, then Voice-over.
3. Under Type a voice-over, type the words. Below the box you'll see how many characters it is, about how many seconds it takes to say, and the price.
4. Pick a voice. Press ▶ on any voice to hear it read the start of your words (up to the first 100 characters). Playing a voice uses no credits. Use All, Female, Male or the accent list to narrow the voices down.
5. Click Make the voice-over. The price shows on the button.

The voice-over appears at once on a Voice-over track at the playhead, striped and labelled ⟳ Making your voice-over…. It usually takes a few seconds. You can keep editing, and it plays as soon as it's done. It's saved to your Library too, named after the voice and the first words, like Voice-over · Matilda · Our new house blend…. It's captioned with your clips, and music is lowered under it like any voice-over.

Price: 1 credit for every 40 characters, at least 2 (a 30-second voice-over is about 12). You're only charged when it's ready.

Change the words or the voice

Select the voice-over on the timeline. Its settings start with Words: the words it says and the Voice.

1. Fix the words (a typo, a price, a name) or pick another voice. ▶ plays the new voice reading your words, for no credits.
2. Click Make it again with these words. The price shows on the button.

The new take replaces the old one in the same place. ⌘Z brings the old take back. If the new take can't be made, the old one stays, nothing is charged, and a note says why.

Good to know:

• A voice-over can be up to 2,000 characters, about 2 minutes. For longer narration, make it in parts.
• Write numbers, prices and names the way they should be said.
• Clean up voice isn't needed for an AI voice; it's recorded clean.
• If a voice-over can't be made, it's taken off the timeline, nothing is charged, and a note says why. Try again, or pick another voice.

Record a voice-over

Use this to narrate while you watch the video.

1. Move the playhead to where the voice-over should start.
2. Click Audio, then Voice-over.
3. Click Record from the playhead. The first time, your browser asks to use the microphone. Allow it.
4. After a 3-second countdown, the video plays while you speak. A timer and a level meter show it's recording.
5. Click Stop.

The take appears on a Voice-over track where you started, and it saves to your Library in the background. You can also add an earlier recording from Your recordings.

Good to know:

• Wear headphones so the video's sound stays out of your recording.
• A take can be up to 10 minutes long.
• If a take doesn't upload, it's kept in the panel. Choose Send the take again when you're ready.
• If the microphone is blocked, allow it from the icon in your browser's address bar, then try again.

Change a sound's settings

Select a sound on the timeline to see its Sound settings.

• Kind: Music, Voice or Effect.
• Lower under speech: for music and effects, turns the sound down while someone talks.
• Volume: up to 200%.
• Starts at and Ends at: which part plays.
• Fade in and Fade out.

A clip's own sound has its settings on the clip's Audio tab: Mute, Even loudness (evens out quiet and loud parts so the sound sits at a steady level, applied when you publish), Clean up voice, Volume (up to 400% for quiet recordings), and Fade in and Fade out.

Good to know: the preview plays volume up to 100% and only approximates the lowering under speech. Both are applied exactly when you publish.

Clean up a voice (no credits)

Use this to reduce hum, traffic or room echo behind someone speaking.

• On a clip: select it, open its Audio tab and switch on Clean up voice.
• On a voice-over: select it. With Kind set to Voice, switch on Clean up voice.

It uses no credits. The preview plays the original sound, and the cleaned sound is used when you publish.

Studio voice

Use this for recordings Clean up voice can't rescue. It keeps only the voice and strips out wind, traffic, chatter or music in the room.

1. Select the clip and open its Edit tab. Studio voice is under AI tools for this clip. For a voice-over, it's in the sound's settings.
2. Click Clean with Studio voice. The price shows on the button before you start, and you're only charged when it's ready.
3. Keep editing. The original plays until the cleaned voice is ready, and the clip is marked ⟳ Cleaning up the voice… on the timeline.

When it's done, the settings say On: this plays the cleaned voice, and the clip's name starts with ✦. Use original goes back to the original at any time.

Good to know:

• It cleans only the part of the recording you're using, from 1 second up to 15 minutes at a time. For a longer recording, split it and clean each part.
• If you lengthen the clip afterwards, the new part plays silent until you choose Clean again.
• If it can't be done, the clip goes back to the original and nothing is charged.

Remove ums and pauses

Use this to tighten talking clips by cutting filler words like "um" and "uh", and shortening long silences.

1. Click Remove ums and pauses in the toolbar above the timeline to tidy the whole video. For one clip, use Remove ums and pauses under AI tools for this clip in its settings.
2. A note says it's listening, usually a few seconds per minute of speech.
3. When it's done, the note says how many ums and pauses were removed and how many seconds that saved.

It uses no credits, and the whole change is one step, so ⌘Z undoes it.

What it changes:

• Pauses longer than 0.6 seconds are shortened, leaving a natural gap of a quarter of a second.
• Cuts are only made where nobody is talking. A voice-over speaking over a pause keeps it.
• Each cut is visible: talking clips and voice-overs become back-to-back pieces, and everything after moves up to close the gap.
• Things that span a cut, like music, a logo or text, just get shorter.
• Captions follow. Removed fillers are listed struck through in the Captions panel.

Good to know: ums can only be found in clips transcribed since 27 September 2026. Older clips have their pauses shortened, and the note says so.

Related: Getting started · Cutting and arranging clips · Text, logos, stickers, templates and captions · AI tools for video · Publishing and exporting · Editing on your phone · Troubleshooting