# Canva text to speech with ElevenLabs voices

Updated 2026-11-06

Canva text to speech only speaks with Canva's own voices, so an ElevenLabs voice has to arrive as a finished audio file. Generate the voiceover in ElevenLabs, download it, then upload it from Canva's Uploads tab and place it on your design. Canva takes audio files up to 250MB, far more than a narrated deck will ever need. ElevenLabs' own Canva guide says the two products can't be integrated directly. That leaves this round trip as the method, and once the script is final it takes a couple of minutes.

## Can Canva text to speech use ElevenLabs voices?

Canva's text to speech can't use ElevenLabs voices at all. The tool is called AI Voice, and you'll find it in the editor under Elements, then Audio, as Generate AI Voice, or through the Apps tab. Type the script, choose a language and a voice style, and Canva puts the generated track on the design's audio timeline. For an internal deck nobody will remember next week, fine.

The voice list belongs to Canva, though. Nothing from the ElevenLabs Voice Library shows up in it. Canva's help center has pages for third-party voice apps like AiVOOV and Murf AI. I looked for an ElevenLabs app in the Apps Marketplace and came up empty, and ElevenLabs' Canva guide (updated October 2026) still walks through downloading and uploading. Check the Apps tab yourself before you treat that as settled. Marketplaces move faster than help pages.

## How do I get an ElevenLabs voiceover into Canva?

Six steps. The first one saves the most time.

1. **Finish the script first.** Each edit after generation means another take and another upload, so time the words against your slides or scenes before you start. My [words per minute calculator](https://voicefieldguide.com/tools/words-per-minute-calculator) converts a target length into a word budget.
2. **Generate it in ElevenLabs.** Open Text to Speech, paste the script, pick the voice from your voices and set the model to Eleven v4. To add one of mine, open the share link on its voice page while you're signed in, and it lands in My Voices.
3. **Download the audio.** A fresh take has its download button at the bottom right. The history panel on the right gives you MP3 or WAV, and its Advanced option adds M4A and FLAC. Canva's list of audio formats is AAC, M4A, MP3, OGG, WMA, WAV and WEBM, at up to 250MB per file. FLAC isn't on it, so skip that one.
4. **Upload it to Canva.** Open Uploads in the editor's side panel and choose Upload files. Dragging the file from your computer onto that tab works too.
5. **Place it.** Click the uploaded file to add it to the design. On a video design it sits on the timeline, where you can move it, trim it and change its volume.
6. **Preview it.** Canva's help suggests Present for checking narration timing. I also play the export once through a phone speaker, because that is where a muddy mix gives itself away.

## Should a Canva presentation get one audio file or one per slide?

Give each slide its own file. Decks get reordered, usually the night before. One long voiceover freezes the original order, and fixing that in Canva means cutting the audio at exactly the right breath, which is fiddly and rarely clean. With a file per slide, the slide and its narration move as a pair.

Canva allows up to 50 audio tracks in one design. I have never sat through a deck that needed more.

Name each download with its slide number before you upload it. The Uploads tab fills quickly, and a dozen files with near-identical timestamps are miserable to sort.

## What should I change in ElevenLabs when a take runs long?

Change the words. Eleven v4 has two voice settings, Stability and Similarity, and ElevenLabs' v4 documentation says the Style and Speed sliders aren't available on v4. There's no control that squeezes a read. When a slide's narration runs four seconds over, cut a sentence and generate again. I'd do that even if a speed slider existed. Sped-up voiceovers sound rushed, and viewers notice it even when they couldn't tell you what bothered them.

Stability is the one setting worth touching for a deck. Low values give a looser, more varied read. High values hold the delivery close to a fixed baseline, which is what twelve separate slide files need if they're going to sound like one recording session. I raise it a little and keep the number identical for every slide.

Pacing on v4 comes from punctuation. Break tags do not work on it, so the pause before an important number has to come from a full stop, a line break or an ellipsis. My guide to [writing a voiceover script](https://voicefieldguide.com/guides/how-to-write-a-voiceover-script) goes further into that.

## Can I make a Canva voiceover from my phone?

Yes, with one extra step. The ElevenLabs mobile app syncs with your web account, so any voice you added on the web is already waiting there. The app can also export clips into other apps. ElevenLabs names CapCut, iMovie and Instagram as examples, and Canva isn't one of them. On a phone I save the file to the device, open Canva and upload it from the Uploads tab like any other audio.

## Which of my voices work best for Canva voiceovers?

Most Canva voiceovers end up as presentations, explainers or training clips. Those want a steady voice.

**Dexter** is my first pick for onboarding decks and product walkthroughs. It doesn't hurry the bit where people follow along, and that counts when the viewer has your product open in another tab. Its weak spot is excitement. Dexter was built for clarity, so a launch-day hype slide comes out sounding polite. Hear it on [Dexter for support and onboarding](https://voicefieldguide.com/elevenlabs-voices/dexter-customer-support-pro/customer-support-and-onboarding).

**Heisenberg** suits lesson slides and explainer videos where each point needs a moment to settle. It sits low and leaves space after every line. That space costs a little time, and under a minute a quicker voice like Reel fits a few more words onto the same slide. On a tight social cut I'd choose something else. See [Heisenberg for explainers](https://voicefieldguide.com/elevenlabs-voices/heisenberg-polished-and-clear/explainers).

For these sample reads I'd use [Dexter](https://voicefieldguide.com/elevenlabs-voices/dexter-customer-support-pro) for a screen recording, [Heisenberg](https://voicefieldguide.com/elevenlabs-voices/heisenberg-polished-and-clear) for a longer lesson, and [Reel](https://voicefieldguide.com/elevenlabs-voices/reel-conversational-voice-for-short-content) for a short social design. Match the voice to the time each slide needs.

## Audition details

These picks compare my fifteen voices. I own them, and ElevenLabs may pay me when you use them.

### Dexter: Screen recordings

Steady support delivery leaves space for a viewer to find each control. Hype copy will sound too polite.

- Voice ID: `Smxkoz0xiOoHo5WcSskf`
- Voice page: https://voicefieldguide.com/elevenlabs-voices/dexter-customer-support-pro
- Add to My Voices: https://elevenlabs.io/app/voice-lab/share/2c79c9f85b01099b7eb10d1340f7f6d43d132db92fe0ede23cd10eef2678f8f1/Smxkoz0xiOoHo5WcSskf
- This guide's Eleven v4 clip: https://voicefieldguide.com/audio/v4/guide--canva-text-to-speech--dexter-customer-support-pro.mp3

Sample line: Open the project, then choose Uploads. Bring in the finished voiceover and place it under the first slide. Play the transition once before you move on; the sentence should finish before the picture changes.

### Heisenberg: Longer lessons

Even educational delivery fits a lesson with several slides. Leave space between slides for the pauses after each line.

- Voice ID: `iEBOK9alpKauGRvBSsFi`
- Voice page: https://voicefieldguide.com/elevenlabs-voices/heisenberg-polished-and-clear
- Add to My Voices: https://elevenlabs.io/app/voice-lab/share/e0464333b21941625fcdd28b2ce050bc74911f6bd69cec0d6861a8bd9e24032b/iEBOK9alpKauGRvBSsFi
- This guide's Eleven v4 clip: https://voicefieldguide.com/audio/v4/guide--canva-text-to-speech--heisenberg-polished-and-clear.mp3

Sample line: Leave this diagram on screen for a moment. Follow the line from the first box to the second, then read the label beside it. That small connection explains why the last step depends on everything before it.

### Reel: Short social designs

Direct conversational delivery suits a short captioned design. It will flatten a dramatic reveal.

- Voice ID: `JKbdwi8BFwQlr1n3fwoT`
- Voice page: https://voicefieldguide.com/elevenlabs-voices/reel-conversational-voice-for-short-content
- Add to My Voices: https://elevenlabs.io/app/voice-lab/share/f7c9d8f6f9db18940ec9cf2c2370d142e3fd94619897fffe57a9e786cdd984dd/JKbdwi8BFwQlr1n3fwoT
- This guide's Eleven v4 clip: https://voicefieldguide.com/audio/v4/guide--canva-text-to-speech--reel-conversational-voice-for-short-content.mp3

Sample line: Here's the part I'd put on the first slide. Show the result, say what changed, and leave the instructions for the next frame. You've got one idea to land before the viewer swipes. Give it the whole sentence.

## Sources

- [Canva Help Center: Turn text into voiceovers using Canva AI voice](https://www.canva.com/help/canva-ai-voice/)
- [Canva Help Center: Upload media from your device](https://www.canva.com/help/upload-media/)
- [Canva Help Center: Check file upload formats and requirements](https://www.canva.com/help/upload-formats-requirements/)
- [Canva Help Center: Add audio tracks and sound effects](https://www.canva.com/help/stock-music/)
- [ElevenLabs: Guide to ElevenLabs text to speech with Canva](https://elevenlabs.io/blog/guide-to-elevenlabs-text-to-speech-with-canva)
- [ElevenLabs docs: Text to Speech product guide](https://elevenlabs.io/docs/eleven-creative/playground/text-to-speech)
- [ElevenLabs docs: Eleven v4](https://elevenlabs.io/docs/overview/capabilities/text-to-speech/eleven-v4)
- [ElevenLabs Help Center: What is the ElevenLabs App?](https://elevenlabs.io/docs/help-center/product/mobile-apps/eleven-labs-for-ios-and-android/what-is-the-eleven-labs-app)

Source page: https://voicefieldguide.com/guides/canva-text-to-speech
