How to use the text to speech reader
- Type or paste your text into Text to speak, or press Read clipboard. The box opens with a sample sentence so you can try it at once.
- Pick a Voice. The list comes from your device and browser, so it differs between Windows, macOS, Android and iOS. A voice marked (online) is generated by a remote service.
- Set Speed (0.5ร to 2ร), Pitch (0.5 to 2) and Volume.
- Press Play. Use Pause, Resume and Stop as you would on a media player. The Live preview highlights the sentence being read and, where your browser reports it, the current word.
How it works
The page calls the speechSynthesis interface of the Web Speech API, so the voices, accents and quality are your browser's, not ours. Long text is cut at sentence breaks into pieces of about 180 characters (fewer at slow speeds) and queued one piece at a time. Some browsers stop a single very long utterance part-way through, and short pieces avoid that.
The speaking-time estimate is the word count divided by 150 words per minute times your speed setting, a common rule of thumb for an average speaking pace. Real voices run a little faster or slower.
Worked example with the sample text
The sample sentence is 30 words and 177 characters, so the counter reads 177 / 100,000 characters. At 1.0ร the estimate is 30 รท 150 minutes, which is 12 seconds. Move Speed to 2.0ร and it halves to 6 seconds. At 0.5ร it doubles to 24 seconds.
Limits and edge cases
- The browser's speech engine cannot save its output, so there is no audio file to download here. For an MP3, use TextSpeakPro.
- Text is limited to 100,000 characters per page load. Longer text is cut off by the box.
- Some voices, especially online ones, do not report word positions. Then only the sentence being read is highlighted.
- Pronunciation of numbers, abbreviations and names is decided by the voice. Spell a word the way it sounds if it comes out wrong.
- If no voice is listed, your device has none installed. Add one in your system's language or accessibility settings.
- To check how long your text takes to read silently, use the Reading Time Estimator. To count its words first, use the Word Counter.
Frequently asked questions
Why do I see different voices on my phone and my laptop?
Each operating system and browser ships its own voice packs. Windows, macOS, Android and iOS all differ, and Chrome or Edge may add online voices on top. Installing another language pack in your system settings adds more voices to the list.
Can I download the text to speech audio as an MP3?
Not from this page. The Web Speech API plays sound through your speakers but gives the page no audio file. If you need a downloadable MP3, TextSpeakPro turns text into natural AI voice audio files.
Is my text sent to a server?
This page never uploads your text. Whether your browser's speech engine does depends on the voice: a voice without (online) next to its name runs on your device, while a voice marked (online) is generated by your browser maker's service. Pick an on-device voice for private text.
Why does the highlight not follow every word?
Word highlighting depends on boundary events from the voice. Chrome and Edge send them for most voices, Safari and Firefox less reliably, and many online voices send none. In that case the current sentence is still highlighted.
Is there a limit on how much text it can read?
The box accepts up to 100,000 characters, roughly 16,000 words. The text is read in short pieces, so long articles play through without stopping part-way.
Sources: MDN, SpeechSynthesis, MDN, SpeechSynthesisVoice.localService and the Web Speech API specification.