How to Use Chinese Text to Speech: Voices, Pinyin, Speed, PDF and Chrome

What Chinese text to speech does, and what it does not

It reads Chinese text aloud in Mandarin using a generated voice. It does not listen to you, score your pronunciation, translate English into speech automatically, or imitate regional accents.

A text to speech tool takes written Chinese and produces audio. The reader here offers one female and one male Mandarin voice, adjustable speed, a pinyin display above the characters, and a short English translation for each sentence when your Chrome version supports it. The audio is generated online, so a connection is required.

Several things people search for are different products. A speech-to-text tool converts your voice into characters. An accent generator changes how an English voice sounds. A pronunciation checker grades your speech. This reader does none of those as its main job, and it does not promise native-app features, Cantonese voices, voice cloning, or offline playback.

Web reader or Chrome extension: which one to use

Use the web reader when you have text to paste, a document to upload, or an MP3 to save. Use the Chrome extension when the Chinese is already on a web page you are reading.

The two products share the same voices and the same daily allowance, but they are built for different moments. The extension keeps a reading panel beside the page, so you never copy and paste. The web reader adds file import and export.

Prepare the text before you press play

Clean, punctuated text sounds better. Remove menus and page numbers, keep Chinese punctuation, and test a short passage before generating a long one.

Voices follow punctuation. A full stop or comma gives the voice a natural place to pause, while a long unpunctuated string of characters tends to sound rushed. If you copy from a web page, remove navigation links, advertisements, and repeated headings first.

The text should be Chinese characters. Pinyin typed in Latin letters is not a substitute for characters, and an English sentence will not be spoken as a Chinese translation. The OCR used for images recognises simplified Chinese and English, so a photo of traditional text may need to be retyped or checked carefully.

Numbers, dates, names, and abbreviations are the usual trouble spots. Listen to a sentence containing them before you rely on a long recording, and rewrite them in words if the result is not what you intended.

Choose a voice

You can choose a female or a male Mandarin voice. Pick the one you find easier to follow, and test it on the same sentence before committing to a long text.

The choice is a gender choice, not a catalogue of named speakers, regional accents, or characters. Because the choice is simple, a good habit is to play the same two or three sentences with each voice once and keep whichever one you can follow more comfortably. Keep that choice for a study session so you are not comparing voices when you should be comparing sentences.

If you are practising for real conversations, listen to both: hearing a different voice for the same sentence helps you recognise the words themselves, not one speaker.

Tones and pinyin: how to listen on purpose

Turn on pinyin, play one sentence at a time, and say it aloud after the voice. Mandarin has four tones plus a neutral tone, and a syllable with the wrong tone can be a different word.

Pinyin shows the syllable and the tone mark above each character. Listening while watching the marks connects what you hear with what you read, which is the main benefit of using a reader instead of plain audio.

• Play the sentence once without looking at the pinyin and guess the tones.
• Play it again with pinyin on and compare the marks to what you heard.
• Slow it down to 0.75x or 0.5x and repeat it aloud, copying the pitch movement.
• Return to normal speed and say the whole sentence with the recording.

Set the speed

The reader offers 0.5x, 0.75x, 1x, 1.25x, 1.5x and 2x. Use 0.5x–0.75x to study a hard sentence, 1x to practise, and 1.25x–1.5x to skim familiar material.

Slow speech makes it easier to separate syllables, but it can make rhythm sound unnatural, so finish a study sequence at normal speed. Fast speed is useful when you already understand most of a text and only want to follow it.

On the website, speed applies to live playback. A downloaded MP3 is generated with the speed you selected, so check the rate before you export and listen to the file afterwards.

Read PDFs, documents, images and web pages

On the website, you can import PDF and DOCX files and recognise text in images. In Chrome, the extension reads selected text or the whole page.

For a PDF or DOCX, upload the file and read the extracted text before generating audio. If a PDF is a scan with no real text layer, the reader falls back to image recognition, which is slower and can make character errors. Always scan the result for obviously wrong characters.

In Chrome, select a passage, right-click and choose Listen with Chinese TTS, or click the toolbar icon to read the page. The panel also works in Google Docs, Slides, Drive, Gmail, Outlook, Kindle samples, ChatGPT and Claude.ai. Chrome restricts extensions on internal pages and some protected destinations, so test a normal public article before assuming something is broken.

Check meaning with sentence translation

If you use Chrome 138 or newer, you can turn on short English translations that are generated on your device. They help you check meaning but they are not a substitute for a translator.

The translation runs through the translator built into Chrome, from Chinese to English. It may download a language model the first time and may be unavailable in your browser. Treat it as a quick check, not a final translation, and compare it with the characters and the audio.

Limits, privacy and cost

Without signing in you can generate up to 3,000 characters per day, and up to 5,000 characters per day with a Google account. Speech is generated online.

Because the speech is generated by a server, your text is sent to produce the audio. Do not process confidential documents unless you have read the privacy policy and are comfortable with that. Split a long document into sections and process what you need, because repeated requests do not bypass the daily allowance.

There is no offline mode, and there is no native iOS or Android app. If a store listing says otherwise under a similar name, it is a different product.

Simplified, traditional, numbers and names

Test a short sample first. The reader and its image recognition are built around simplified Chinese, and numbers, dates and personal names are the places a voice is most likely to surprise you.

Characters that look the same in simplified and traditional text are read the same, but the tool is designed around simplified text, and image recognition uses a simplified Chinese model. If you work with traditional text, play a representative sentence before you rely on a long passage, and check that rare characters are read as you expect.

Numbers can be read as quantities, as digits one by one, or as years, depending on context. Phone numbers, product codes and years are worth checking. If a name is read with a surprising tone, test it inside a short sentence, because many words depend on neighbouring characters for their reading.

Mixed Chinese and English is common in real text. Short English words and abbreviations can sound unnatural in a Mandarin voice, so listen to those sentences first. If the result matters, write the term in Chinese or spell it out.

Who uses it, and for what

Learners use it to hear and repeat sentences, teachers use it to prepare listening material, translators and editors use it to check how a text sounds, and many readers simply listen to long articles.

Learners get the most from short, repeated passages with pinyin on. Teachers can paste a reading text, set the speed for the class, and export an MP3 on the website for a worksheet or lesson. Translators and editors listen to a draft to catch awkward phrasing and missing punctuation, since a sentence that is hard to read aloud is often hard to read silently too.

For long reading, the panel in Chrome keeps the page and the audio together: you follow the highlighted sentence, pause with Alt+P, and jump between sentences with Alt+Left and Alt+Right. Because the same voices are used on the website, a passage you listened to in the browser can later be prepared as an MP3 without changing how it sounds.

If something does not work

Most problems are a blocked page, text that is an image, a daily limit, or a connection issue. Work through them in that order.

If the panel does not open, reload the tab, check that the extension is enabled at chrome://extensions, and try a public article, because Chrome restricts extensions on internal pages and some protected destinations. If the page swaps its content while you are listening, select the passage again.

If no Chinese is found, copy a sentence manually to see whether the page exposes real text. Text inside images, canvases or embedded documents may not be readable by the extension; upload the image or document to the website reader instead.

If playback stops with a limit message, you have used the daily allowance. Signing in with a Google account raises it from 3,000 to 5,000 characters. If audio fails entirely, check your connection, because both playback and MP3 export require the service.

Ten common mistakes

Most disappointing results come from the input, not the voice: unclean text, no punctuation, listening only at one speed, or expecting features the tool does not offer.

• Pasting a whole web page, including menus, instead of the article.
• Using text with no punctuation and expecting natural phrasing.
• Typing pinyin instead of characters.
• Trusting OCR on a blurry photo without reading the characters back.
• Generating a long MP3 before testing one sentence.
• Practising only at 2x or only at 0.5x.
• Listening without looking at pinyin when you are still learning tones.
• Expecting the tool to assess your pronunciation.
• Searching for “accent generator” and expecting an English voice to sound Chinese here.
• Assuming the extension can export MP3.

A five-minute practice routine

Choose three sentences, listen at normal speed, slow down, repeat aloud, then try to say them again without the text.

Take three sentences you partly understand: 今天早上我喝了茶。下午我要去图书馆。晚上我会复习中文。 Play them once without pinyin. Play them again with pinyin on and note any tone you got wrong. Slow them to 0.75x, repeat each sentence aloud, then play at 1x and speak with the voice. The next day change one noun or verb and run the routine again. A small, repeated exercise helps more than a long recording you only play in the background.

Frequently asked questions

Is there a Chinese text to speech that shows pinyin?

Yes. The reader and the Chrome extension both display pinyin above the characters, and you can turn it off in the panel.

Which voices are available?

A female and a male Mandarin voice. There is no choice of named speakers, regional accents, or Cantonese.

Can I download Chinese text to speech as an MP3?

Yes, on the website. The Chrome extension does not export MP3; its download button opens the website.

What is the daily limit?

Up to 3,000 characters per day without signing in and up to 5,000 characters per day with a Google account.

Does it work offline?

No. Speech is generated online. Only the optional sentence translation runs on your device.

Can it read a PDF?

On the website you can import PDF and DOCX files. Scanned PDFs fall back to image recognition, which can make character errors, so check the extracted text.

Does it check my pronunciation?

No. It plays a Mandarin voice for you to listen to and imitate; it does not record or grade your speech.

Is there a mobile app?

No native iOS or Android app is offered. You can use the website in a mobile browser, and the extension works in desktop Chrome.

Related reading

• How to Pronounce Chinese: Hear Words and Sentences with Pinyin
• Read Chinese Text Aloud: Web Pages, PDFs, Docs and Email
• Chinese Reader with Pinyin: Read Mandarin Aloud
• Chinese PDF Text to Speech: Read Documents Aloud
• Chinese Text to Speech MP3: Text to Voice and Audio Download
• Chinese Text to Speech Chrome Extension: Read Aloud

Return to the online reader. Review the privacy policy and terms before using sensitive or published material.