About Text to Speech
Text to speech runs on the browser's built-in Web Speech Synthesis API. Your text is never uploaded to OLTool servers, though cloud-backed voices may send it to the browser or OS vendor for synthesis. Pick from the languages and voices installed on your system; rate and pitch each range from 0.5 to 2 (0.1 steps); while speaking, the current word is highlighted, with pause, resume and stop available.
Available voices are decided by the operating system and browser: SAPI on Windows, system voices on macOS, and often speech-dispatcher on Linux. An empty list or no Chinese voice usually means a missing system speech pack, not a page failure. This tool submits the whole text as one utterance and does not split it automatically; browsers may silently stop on very long single utterances (common in Chrome), so read long documents paragraph by paragraph. Pronunciation of polyphones, numbers and abbreviations is up to the engine and cannot be specified here.
How It Works Internally
Speech synthesis segments text into linguistic units and renders audio. Readability is driven mostly by rate and sentence segmentation, with pitch carrying emotion; long text must be submitted in chunks split after full stops, since cutting inside a number-plus-unit or an English word causes mispronunciation. Markup support varies by engine and built-in browser voices largely ignore it.
Splitting "3.5 kilograms" in the middleMay be read as two fragments or with a wrong unit — split after sentence ends, never inside a value and its unitMeasured on This Page
| Availability check | speechSynthesis is present on load, and pressing Speak with no voices installed raises no exception (no pageerror) |
|---|---|
| Editable input | Text plus rate and pitch are adjustable before speaking |
| Voice list | A VOICE section lists the voices the browser exposes (count depends on installed system language packs) |
| Fully client-side | Speech is synthesised by the browser engine; the text never reaches a server |
The bottleneck in a TTS tool sits outside the page (0 errors measured in headless). Available voices come from the operating system — SAPI on Windows, system voices on macOS, often an empty list on Linux unless speech-dispatcher is installed, so 'no Chinese voice' is a system configuration issue rather than a bug. Headless browsers and some accessibility environments have no audio output, where synthesis finishes or fails silently; measured here without an exception. Voice naming is also not unified across browsers, so the same voice label may differ. Play a short sample before reading long text, to confirm voice and rate before committing.
How to Use
- Open "Text to Speech"
- Paste the text you want read aloud
- Choose a language and a specific voice
- Drag the sliders to set rate and pitch
- Click "Speak", then pause or stop anytime
Use Cases
- Assistive reading — Convert web articles, ebooks or documents to speech for hands-free listening.
- Language learning — Input foreign language text and learn correct pronunciation through speech synthesis.
- Proofreading — Listen to your text being read aloud to catch errors and awkward phrasing more easily.
- Accessibility — Provide text-to-speech output for visually impaired or dyslexic readers.
- Voiceover drafts — Turn slide or video scripts into speech for quick narration drafts.
FAQ
Does it need internet?
Most system voices synthesise offline, so no network is required; some cloud-backed voices send the text to the provider for synthesis and usually fail when offline.
What languages are supported?
All languages available in your browser's built-in voice library, typically including Mandarin Chinese, English, Japanese, Korean, French, German and more.
Can I download as audio file?
The Web Speech Synthesis API does not support direct audio export. Use system recording tools to capture the output.
Why is the voice list empty or silent?
The speech engine for that language is usually missing. Install the language pack in your OS text-to-speech settings, then refresh the page.
Can it read long texts?
This tool submits the whole text as one utterance and does not split it. Very long text may stop silently part-way in some browsers (Chrome especially), so read it paragraph by paragraph or shorten each pass.
Do rate changes apply immediately?
No. Rate, pitch and voice apply only to the next utterance — after changing them, press stop and then Speak again; current playback does not follow the sliders. Pause, resume and stop controls are provided.
Can the generated speech be used commercially?
It depends on the engine's licence: the browser's built-in Web Speech API uses the system TTS whose commercial terms vary by OS, while cloud TTS is billed and licensed explicitly. The tool adds no restrictions — check the engine's terms before commercial use.
Polyphonic Chinese characters are mispronounced — what can I do?
Built-in TTS accuracy depends on the system engine and the tool cannot force a reading. Workarounds: rephrase to avoid ambiguity or add punctuation to fix rhythm. For exact pronunciation use a cloud engine that supports SSML phoneme tags.