ElevenLabs is an AI audio platform that can read written text aloud in a chosen voice. In this guide, you'll create a short voiceover for a personal video introduction, listen to it, make one improvement, and save the audio file. You only need a short script and a web browser. The goal is one finished recording, not learning every model or advanced setting.
Want to compare features, subscriptions, APIs and commercial licensing first? Read our ElevenLabs features and pricing overview before choosing a plan.
Before you start
Open ElevenLabs and create or sign in to an account. ElevenLabs documents a free tier with Text to Speech and a monthly usage allowance, so a short practice clip can be made without a paid subscription. The amount of remaining credits and available voices may vary; check your account before generating. The free plan is for personal, non-commercial use and requires attribution under ElevenLabs' published terms. If you intend to monetize a voiceover, review the current licensing information and choose an eligible paid plan first.
Have a 30–60-word script ready. Avoid uploading private information or someone else's copyrighted material without permission. You do not need an API key, microphone, or voice clone for this exercise.
How to use ElevenLabs Text to Speech
Step 1: Open Text to Speech
In the ElevenLabs web app, find Text to Speech in the sidebar. This opens the text-to-speech workspace. ElevenLabs calls this area its Text to Speech playground in the official quickstart. The layout may adapt to your screen size.
Step 2: Paste your script
Locate the text input box and paste the sample below, or use your own short introduction. Keep your first attempt simple so you can judge the pacing and pronunciation. Punctuation matters: commas can suggest short pauses, and full stops separate ideas.
Welcome to my little creative studio. Today, I'll show you how one simple idea becomes a finished project. We'll start with a plan, try something new, and improve it together. Thanks for joining me.
Step 3: Choose a voice
Use the voice selector to pick a voice that fits the message. Preview available voices when that option appears. For this friendly introduction, try a conversational voice rather than an exaggerated character performance. The Voice Library guide explains how to explore voices. You do not need to clone a real person's voice.
Step 4: Keep the settings simple
Use the default model and voice settings for the first generation. Model choices and controls can change; ElevenLabs currently documents models including Eleven v4 and earlier versions. If your selected model offers Stability and Similarity, leave them at their starting values until you've heard the first result. Not every model supports the same controls. Avoid changing several settings simultaneously because you won't know which change affected the voice.
Step 5: Generate and listen
Click Generate or Generate speech, as shown in your interface. The platform processes the text and produces an audio clip you can play back. Listen from beginning to end. Check that every word is present, the tone fits your message, and any names or abbreviations sound correct. Generation consumes account credits, even if you don't download the result, so review the text before clicking.
Step 6: Make one improvement
If the narration feels rushed, first try adjusting the punctuation: add a full stop where a sentence needs a clearer ending. If a name sounds wrong, spell it phonetically or replace a difficult abbreviation with the words it stands for. Regenerate only after deciding what to fix. Depending on the model and current account rules, some regeneration options may be available, but do not assume they are free.
Step 7: Download the finished voiceover
After generation, use the download button at the bottom-right of the generated result. ElevenLabs also lets you retrieve earlier clips from the History tab in Text to Speech. Its download documentation describes MP3 and WAV options in History, with additional formats under Advanced. For a basic personal project, MP3 is an easy starting point. Save the file with a recognizable name, such as studio-intro-voiceover.mp3, and play it locally to confirm the download.
Example input and expected result
The example script above is approximately 35 words. You should expect one short spoken introduction, delivered by the voice you selected. The exact sound, duration, emphasis, and pauses depend on the voice, model, and settings. A successful result is an audio file that reads your complete script clearly and can be played outside ElevenLabs.
Three tips for better results
- Write for listening, not reading. Replace long, complex sentences with shorter ones and read your script aloud once before generating.
- Spell out ambiguous numbers and acronyms. Write “twenty twenty-six” instead of “2026” when that is the pronunciation you want. ElevenLabs recommends writing out numbers and symbols when pronunciation matters.
- Change one thing at a time. Compare the first recording with a version that changes only the script, voice, or one supported setting. This makes improvements easier to judge and avoids wasting credits.
Common beginner problems
The voice mispronounces a name
Try a phonetic spelling or a more familiar equivalent. For a recurring project, investigate pronunciation controls available to your chosen model rather than regenerating blindly.
Generation uses credits
ElevenLabs says generating audio consumes credits, even before downloading. Keep early experiments short and check the remaining allowance shown in your account. Avoid repeated clicks while waiting for a result.
You can't find your earlier recording
Return to Text to Speech and open History. On narrow screens, ElevenLabs documents a History icon near the Generate speech control. Use the download action on the saved generation.
What to try next
Now that you have a first recording, create a second take with a different voice, try a slightly longer script, or combine your MP3 with a personal video editor. If you're new to AI tools generally, browse the DPAI AI tools directory and the practical learning guides for complementary workflows. Keep the license restrictions in mind before publishing or selling generated audio.
Frequently asked questions
Can I use ElevenLabs for free?
Yes. ElevenLabs documents a free tier for experimenting with Text to Speech. Check the current allowance and available voices in your account; commercial rights are not included in the free tier.
Do I need an API key?
No. This tutorial uses the browser interface. API keys are for developer integrations, not the basic Text to Speech workflow.
Can I download an MP3?
Yes. The generated result has a download action, and the History area provides MP3 options according to ElevenLabs' documentation.
Can I use the recording in a monetized YouTube video?
Do not assume a free-plan recording has commercial rights. ElevenLabs states that commercial use requires an eligible paid plan and is subject to its terms. Check those terms for your specific use.
What if the voice sounds unnatural?
Try shorter sentences, clearer punctuation, a different voice, or a single supported setting adjustment.