How to Do a Voiceover on Google Slides

Short answer: To do a voiceover on Google Slides, record or generate the narration outside Slides, upload each clip to Google Drive as an .mp3 or .wav file, then add it to a slide with Insert > Audio. Google Slides does not record voiceover clips inside the editor the way PowerPoint can, so you build the narration one slide at a time. Google also offers a slides recording feature that captures your screen and microphone into a single video, but that produces a video file rather than editable per-slide audio.

This guide covers the audio formats Slides accepts, the Drive step that trips up most people, the per-slide workflow that scales to a full deck, and the fixes for the usual playback errors. The steps assume the desktop web app, because inserting audio is a desktop-only action even though audio that has been inserted plays on other platforms.

Why a Google Slides voiceover is a two-app job

Slides leans on Drive for its media. When you choose Insert > Audio, the picker lists only audio files already stored in your Google Drive, and it accepts .mp3 and .wav only. There is no microphone button in the toolbar and no record narration wizard for individual slides. That design choice is why adding a voiceover to Slides is always a two-app job: an audio recorder or a text to speech tool creates the file, and Slides places it.

Google later added a recording feature, but it works differently from a per-slide voiceover. Slides recording captures your presentation with your webcam or microphone and saves the result as a video. Google's help page states that it runs only in Chrome and Edge, may only be available on certain work or school accounts, requires edit access to the deck, allows a 30 minute recording at a time, and stores the file in a My Drive folder called Slides recordings. Its known limits are:

• Recording works only in Chrome and Edge on desktop (you can view and share a recording from any browser).
• The feature may only be available on certain work or school accounts.
• There is a 30 minute limit on every single recording.
• Recordings are saved as videos in the Slides recordings folder in Drive.
• Recordings count against your Drive storage quota, and you cannot create more if you are over quota.

That feature is a good fit when you want one shareable video of a talk. It is the wrong fit when you want narration that lives on each slide and can be edited slide by slide, because it hands you a single video instead of separate clips.

What audio Google Slides accepts, and where it has to live

Google's help is specific: you can add .mp3 and .wav files stored in your Drive, and .wav playback is supported in Chrome, Firefox, Safari, and Microsoft Edge. Anything else is not on the accepted list, including the .m4a files that Apple Voice Memos and QuickTime produce by default, so convert before you upload. The clip also has to sit in the same Google account you are editing the deck with. A file that is clearly in Drive but missing from the picker is almost always in a different account.

Uploading is a separate step from inserting. In Drive, use New > File Upload, or drag the file into a folder. Google notes that an uploaded file takes up space in your Drive even when you upload it into a folder owned by someone else, so a narrated deck with dozens of clips adds up. Keep clip names short and ordered, like slide-01.mp3 and slide-02.mp3, so the picker is easy to scan when a deck grows past ten slides.

Add a voiceover to one slide, step by step

Work one slide at a time. Doing a whole deck in one pass is where timing and naming mistakes creep in. For each slide:

1. Write that slide's narration in the speaker notes, so the script stays attached to the slide it belongs to.
2. Record or generate the audio for the slide as a separate clip, then export it as .mp3 (or .wav).
3. Upload the clip to Drive with New > File Upload, or drag it into a Drive folder.
4. In Slides, select the slide, open Insert > Audio, choose the file, and click Select.
5. A small speaker icon appears on the slide. Hover over it to see the playback controls.
6. With the icon selected, open Format options to set playback, volume, and looping, and to hide the icon or replace it with an image of your choice.
7. Repeat for every slide, then test the deck in Present mode before you share it.

The Format options panel is where a slide's audio stops being just a sound file and becomes part of the presentation. Hiding the speaker icon keeps the slide clean, and replacing it with an image lets a picture or a button start the clip. Those options are documented in Google's own announcement of the feature, which also confirms that inserting audio is a desktop action while playback works on other platforms.

The per-slide workflow that scales to a full deck

There are three realistic ways to lay narration over a deck, and they behave very differently once a viewer starts clicking through. Pick the shape that matches how the deck will be used.

For most narrated decks, one clip per slide is the approach that behaves predictably. Each clip starts when its slide appears and ends when the viewer moves on, so a slow reader cannot pull the audio out of sync with the visuals. Naming matters more than people expect: when slide 12 needs a fix, you want to find slide-12.mp3 in the Drive picker without guessing which take is current. Keep a single master script in the speaker notes so a re-record is quick.

Fixing the usual Google Slides audio problems

Almost every Google Slides audio problem falls into a short list. Check these before you assume the feature is broken:

• The file is missing from the picker: confirm it is .mp3 or .wav and stored in the same Google account you are editing with. Other formats and other accounts do not appear in the list.
• You hear it but viewers do not: this is a Drive sharing setting. Share the audio file, or the folder that holds it, with the same audience as the deck, then reload and test from a second account.
• Autoplay does nothing on the first slide: browsers can block sound that starts without a click, so click once in Present mode when you begin, or set the clip to start on click if you prefer predictable control.
• A clip bleeds into the next slide: turn stop on slide change back on for that clip.
• You cannot trim a clip inside Slides: Google's help documents playback, volume, and looping rather than trimming, so cut the clip in your audio editor and re-upload.

When Google Slides is the wrong tool for the job

Slides is a presentation editor, and its audio features are deliberately light. If you need to trim clips on a timeline, cross-fade music under a voice, stack several voice tracks, or export a finished edit, a dedicated audio or video editor is the better home for that work. If your audience needs narration you speak live, in a webinar or a live class, recording files ahead of time is the wrong shape entirely. If your team is on Windows only, Slides is still fine for the deck, but any Mac-only tool you pair with it will not be available to those teammates.

Be honest about the boundary on the audio side too. A tool that generates speech does not replace a person who needs to respond to questions in real time, and it is not a fit for a live, interactive session. Where generated narration shines is the asynchronous case: a training module, an explainer, or a self-running deck that a viewer opens on their own schedule. That is exactly the case Slides is built for.

Recording or generating the narration on a Mac

On a Mac you have two ways to produce the clips: record your own voice, or generate the narration as speech from text. Recording yourself with QuickTime or Voice Memos is direct, but both save .m4a by default, so you export to .mp3 before the Drive step. Generating the audio skips the recording session and lets you fix a single sentence without a full retake, which matters when a deck has forty slides.

Murmur is a local voice studio for Apple Silicon Macs running macOS 15 or later, and it is one way to produce those clips on a Mac. It runs its speech models on the machine, so narration text and audio are not uploaded to a cloud service and there are no per-character fees. Kokoro is bundled and ships with more than 10 built-in voices, and several of the installable models support cloning a voice from a short reference sample. That is useful when a client wants one consistent narrator across a whole course or deck.

Once the clips exist, the rest is the workflow above: export .mp3, upload to Drive, insert per slide, and set playback. Generate all the clips in one sitting with the same voice so pitch and pace stay consistent across the deck. The Mac text-to-speech workflow covers the capture and export step in detail, and the audio export guide explains which file to keep as your master. This is the same pattern used for local TTS for training videos, where one voice has to carry narration across many slides. If avoiding per-character pricing is the priority, the approaches in AI voiceover without a subscription start from the same local-first idea.

Frequently asked questions