Text to speech for YouTube
Write your script, pick a voice, and download a voiceover plus subtitles for your video.
Recording your own narration means a quiet room, a decent microphone and a lot of retakes. For explainers, tutorials, list videos and faceless channels, a clean synthetic voice is often enough, as long as the pacing fits the edit. This page covers how to write and structure a script so the voiceover drops into your timeline without fighting it: where to put pauses for B-roll, how to split long scripts into takes, which voices suit which kind of video, and how to get subtitles that line up with the audio.
Open the tool → Opens the tool with the American male voice Michael and MP3 as the download format, and fills in an example YouTube intro script if the box is empty.- Free, no limitNo account, no character cap, no watermark.
- Nothing is uploadedYour text is turned into speech on your device and never reaches a server.
- Works offlineAfter the first use the tool keeps working with no connection.
- Audio and subtitlesDownload MP3, WAV, or an SRT file timed to the speech.
Write for the ear, then for the edit
A script that reads well on screen can sound stiff when spoken. Keep sentences short, put the important word near the end of each sentence, and avoid stacking three clauses with commas. Read a paragraph back with the play button before you commit to it; if you stumble mentally, the voice will sound flat there too.
Think about what the viewer sees while each line plays. If a sentence introduces a screen recording or a product shot, the viewer needs a beat to look at it. That is where pauses earn their place, and it is much easier to add them in the text than to stretch audio later in your editor.
Pacing for B-roll and screen recordings
Type [pause 1s] or [pause 2s] wherever you want the narration to breathe, for example right after "Here is what the settings page looks like." You can also use the Pause button to insert 0.5, 1, 2 or 3 seconds at the cursor. A blank line between paragraphs gives a natural short pause, which works well as a gap between sections of a video.
For tutorials, a speed of 0.9 or 1 gives viewers time to follow clicks on screen. For list videos and commentary, 1.1 feels more energetic without sounding rushed. Pick one speed per video and keep it, because changing speed between sections is noticeable.
Split long scripts into takes
For a ten-minute video, it helps to generate each section of the script separately and download each one as its own MP3. Name the files by section, such as 01-intro and 02-setup, and they will sit in order in your editor. If you later rewrite one section, you only replace that one clip instead of re-aligning the whole timeline.
Within a single take, you can still fix mistakes cheaply. After generating, the parts list shows each sentence or two; click a part to hear it, edit its text, and regenerate only that part. Unchanged parts are reused instantly, so fixing a typo in the middle of a long intro takes seconds.
Leave room for music and sound
Background music competes with the voice in the same frequency range, so plan for it. Leave a short pause of one or two seconds before the first line so an intro sting can play, and add pauses where you want the music to swell between chapters. Keep the music low under narration and duck it further under dense sentences.
If you plan to process the audio in your editor with compression or EQ, download WAV instead of MP3 so you start from the uncompressed file. For simple edits, MP3 is fine and smaller to move around. See text to WAV for when the difference matters.
Choosing a voice for your channel
Michael and Fenrir are American male voices that work well for tutorials and explainers. Heart is the most natural-sounding female voice and works for almost any topic. For a calmer, closer sound, such as study or sleep content, try Nicole. Emma and George give a British accent, which some channels use to stand out; more on that in British text to speech.
Each voice has a short listen button in the voice menu, so you can compare them before generating anything. Whichever you choose, keep the same voice across your uploads; viewers come to recognize it.
Frequently asked questions
Can I use the voiceover in monetized videos?
The audio has no watermark and we do not require attribution. Whether a video qualifies for monetization is decided by the platform under its own policies, so check those for your channel.
How do I get subtitles for the voiceover?
Choose SRT as the download format. The timing of each part comes from the real length of the generated audio, so the subtitles line up with the voiceover, and lines are kept to 42 characters and two lines. See text to speech with subtitles for details.
Is there a length limit for my script?
No. There is no character limit and no daily quota. Long scripts are split into short parts, and playback starts while later parts are still being generated.
The voice mispronounces my channel name. What can I do?
Unusual names and brand names are sometimes read wrongly. Spell the word the way it sounds in the script, and regenerate just that part from the parts list.
Can I have two voices talking to each other?
Not in one file. The tool uses one voice per generation. You can generate each speaker's lines separately with different voices and combine the clips in your video editor.
Is my script uploaded anywhere?
No. Everything runs in your browser and the text is never sent to a server. After a one-time download the first time you use it, it also works offline.