Text to speech
Turn text into natural speech with a voice that runs on your device: listen, then save it as MP3 or WAV with subtitles. Free, with no account or uploads.
Your text
Type or paste it, or open a document. It stays on your device.
No words yet
English only for now: the voices read American or British English.
How it is read
The voice runs on your device. The first time, it is downloaded from KeepSimp (116.07 MB) and kept by your browser.
1 is the voice's own pace. From half as fast to twice as fast.
Silence added after each sentence.
Silence added after each paragraph and heading.
MP3 plays everywhere. WAV is lossless and larger.
Each sentence timed exactly, at most two lines of 42 letters and seven seconds.
−16 LUFS with peaks under −1 dBTP, as Apple Podcasts asks of spoken word.
Written into the MP3, for a Markdown or Word file with headings. Apple's apps read another kind of chapter.
Headings are kept, marks and links are read as their words, and code is left out.
- PrivateThe voice runs in this browser. Your text never leaves your device.
- Works offlineAfter the first time, the voice is kept and nothing more is downloaded.
- Hear it at onceIt starts reading while the rest is still being made.
A little help along the way
Follow these steps. Every option is explained underneath. The voice runs in your browser; your text is never sent anywhere.
- Add your textType or paste it into the box, or press Open a document for a text file (.txt), a Markdown file (.md) or a Word file (.docx). Under the box you see how many words it has and about how long it will take to say.
- Say what it is forListen to a document reads a little quicker than talking pace. Video voice-over reads at talking pace, evens the loudness and adds subtitles. Audiobook leaves longer pauses between paragraphs and writes a chapter for each heading.
- Choose a voicePress the play button beside a voice to hear it say a sentence; it is made on your device. There are nine voices, six with an American accent and three with a British one.
- Make speechPress Make speech. The first time, the voice's files are downloaded from KeepSimp and kept by your browser; after that it starts at once. It starts reading while the rest is still being made, and Stop halts it within a sentence.
- DownloadWhen it is ready, the result says how long it is, how many words and which voice. Press Download for the MP3 (or WAV), and the subtitles too when you chose them. Each file is opened again and checked before it is offered.
Every option explained
- Speed
- From half as fast to twice as fast as the voice's own pace. Kokoro speaks faster or slower itself; nothing is stretched afterwards.
- Pauses
- Seconds of silence after each sentence and after each paragraph or heading. They are laid between the sentences afterwards, so changing them does not make the voice speak again.
- File type
- MP3 plays everywhere and is written gaplessly with LAME. WAV is 16-bit and lossless, about ten times larger.
- Subtitles
- SRT or WebVTT, with each sentence at its exact time and its words shared out by length, at most two lines of 42 letters and seven seconds a subtitle.
- Even loudness
- Measured as broadcasters measure it (ITU-R BS.1770) and set to −16 LUFS with peaks under −1 dBTP, the level Apple Podcasts asks of spoken word.
- Chapters
- A chapter for each heading of a Markdown or Word file, written into the MP3 (ID3 chapters). Players such as VLC show them; Apple's apps read another kind.
- Markdown
- Read as Markdown, headings stay headings, links and marks are read as their words, and code is left out. It is turned on when a .md or .docx file is opened.
- What it cannot do
- It reads English only for now, it has no pitch setting, and it cannot copy anyone's voice. It is a computer-made voice and can mispronounce names, numbers and abbreviations.