Text to Speech
Have any text read aloud with an offline English voice you choose, and download exactly that audio as a WAV file. WebToolGrid never receives your text — see "How it works" for how voices are handled.
0 characters
Speed and pitch controls aren't offered for the offline voices, so the audio you hear is exactly what you download.
How it works
Play and Download audio use the same offline neural voice, which runs entirely in your browser, so the file you save sounds like what you heard. Pressing Play generates the audio for your text first, which takes a moment, then plays it. Download audio saves that same audio without generating it again; it is only generated again if you change the text or the voice. The first time you use a voice, it is downloaded from a third-party host and cached by your browser, and each voice's approximate size is shown in the list. The offline voices have no speed or pitch control, so those sliders appear only in the fallback described next.
If the offline voices can't be loaded (for example, with no internet connection the first time), the page uses your browser's built-in speech synthesis (the Web Speech API) for Play, with speed and pitch sliders. Some of those voices are cloud-based (for example, in Chrome or Edge) and send the text to the browser vendor's service. In that mode Download audio is turned off, with the reason shown, because browser voices can't be saved as a file.
Voice availability
The voice list shows the offline English voices this tool loads, each labelled with its country, quality and download size, and it is the same on every device. Only in the fallback does the list come from your operating system and browser, where it varies by device and a website can't add new voices.
What the tool does with your text and the saved file
- Leading and trailing spaces are trimmed, and an empty box shows "Type or paste some text first." The page sets no limit on the length, so a long text simply takes longer to generate.
- The download is a WAV file. Its name comes from the first 30 characters of your text, then -speech.wav: the text "Hello, this is a test of the offline voice." is saved as Hello, this is a test of the o-speech.wav. A full stop within those characters cuts the name short, so Hello. How are you today? is saved as Hello-speech.wav. Characters that are not allowed in file names, such as a slash, colon or question mark, are replaced with a dash.
- Audio is generated once for a given text and voice and reused. Pressing Play and then Download audio does not generate it twice, but changing the text or the voice means new audio is generated.
If the offline voices are not available
The page then uses your browser's own speech synthesis. To avoid a known problem in Chrome where a long utterance can be cut off after about 15 seconds, the text is split at sentence ends (a full stop, exclamation mark, question mark or line break) and sentences are grouped into pieces of up to about 200 characters. A single sentence longer than that is spoken as one piece. This mode shows speed (0.5 to 2) and pitch (0 to 2) sliders, and Download audio is turned off.
Frequently asked questions
Can I download the speech as an audio file?
Yes, as a WAV file. Play and Download audio both use the same offline English voice, which runs inside your browser, so the file sounds like what you heard. The first time you use a voice it is downloaded from a third-party host (roughly 60 MB) and cached by your browser. If the offline voices can't be loaded, Play falls back to your browser's own voices and Download audio is turned off, because those voices can't be saved as a file.
Which voices can I choose, and why do they differ from my browser's?
The list shows the offline English voices, each with its country, quality and approximate download size. They are the same on every device. Only if those voices can't be loaded does the list switch to your browser's own voices, which depend on your operating system and browser.
Is any of my text sent to a server?
Never to WebToolGrid — we don't receive the text you type. The offline voices generate the audio on your device, and Play uses them too, so your text stays on your device. Only if those voices can't be loaded does Play fall back to your browser's own voices, and some of those (for example, some cloud-based voices in Chrome or Edge) send the text to the browser vendor's service.
Why did my downloaded file get a name made of my own words?
The file name is built from the first 30 characters of your text, followed by -speech.wav, so you can tell recordings apart. Forbidden characters are replaced with a dash, and a full stop in those 30 characters ends the name at that point.
Is there a limit on how much text I can enter?
The page does not set one. Longer text takes longer to turn into speech, and very long text may be slow on a phone or an older computer.
Why can't I change the speed of the voice?
The offline voices have no speed or pitch control, so that the audio you hear is exactly what you download. The sliders appear only when the page falls back to your browser's voices.