How to Make a YouTube Voiceover with Free Text-to-Speech

Start with the words, not the voice. Write the way you would explain the topic to a friend: short sentences, one idea at a time, and a clear opening that tells viewers what they will get. Read our guide on writing scripts for text-to-speech for detailed tips on numbers, names and punctuation.

As a rough guide, natural narration runs at about 130 to 160 words per minute. A 5-minute video therefore needs around 650 to 800 words of script.

Choose a voice in the language and accent of your main audience, then test your top candidates with a real paragraph of your script. Explainers and tutorials usually work best with a calm, clear voice at normal speed; short videos can use a brighter voice and a slightly faster pace. Our guide on choosing the right AI voice walks through the process.

Once you have chosen, note the voice name and settings and keep them for the whole channel. A consistent voice becomes part of your channel’s identity.

Split the script by scene or section rather than generating everything at once. In the Studio each request can hold up to 2,000 characters, which fits most sections comfortably. Generating per section has two advantages: you can fix one part without redoing the rest, and each file lines up naturally with a scene in your editor.

Name your files in order (01-intro, 02-problem, 03-step-one…) so they drop into your timeline in the right sequence.

• Use punctuation and pause tags to add breathing room before important points and between steps.
• Leave a short silence at the start and end of each file so cuts in your editor are clean.
• If viewers need to follow along on screen, slow the speed slightly rather than adding long pauses.
• For exact timing to match a visual cut, trim or extend silence in your video editor; that is more precise than adjusting the script.

Import the narration files into your video editor and place them on their own track. Then:

• Keep the voice clearly on top. Background music should sit well below the narration; lower it further (or "duck" it) while the voice is speaking.
• Check overall loudness. Online video platforms normalise loudness. Many creators aim for roughly -14 LUFS integrated for the final mix, but the most important thing is that speech is clear and consistent from start to finish.
• Listen on a phone. Most viewers will hear your video through phone speakers or earbuds, so check the mix there before publishing.
• Add captions. You already have the script, so captions are easy, and they help viewers who watch without sound or who are deaf or hard of hearing.

Platforms update their policies on synthetic media regularly. Before publishing, check the current YouTube rules on disclosing altered or synthetic content and on monetisation. As a baseline:

• Never use a synthetic voice to impersonate a real person or make it seem that someone said something they did not.
• Be open with your audience that the narration is AI-generated when your platform requires it, or when viewers could reasonably be misled.
• Make sure your content adds real value: original scripts, research, visuals and editing. Videos that simply read out copied text are a poor experience for viewers.
• Read our Terms of Use for the rules that apply to audio created on this site.