Chinese audio to text

How do I turn Chinese audio into text?

Choose a recording, drop it on the box or paste it: an MP3, M4A or WAV, a voice message, or a video such as an MP4. Its Mandarin is written out a minute at a time, each line with its time and the pinyin over every word in tone colours. Tap a time to hear that line, tap any word for its meaning, then copy the text or download it as subtitles (SRT).

How long can a recording be?

Up to 10 minutes. Without an account you can try 3 minutes a day; a free account hears 30 minutes a day. For recordings up to an hour, add them to Immersion and press Make a transcript, and the transcript is kept with the recording and reads along as it plays.

How accurate is it?

It is heard by Whisper, OpenAI’s open speech model, which writes clear Mandarin well. Music, noise, people talking over each other and strong accents make it miss words, so read along with the recording: the pinyin over each word makes a wrong character easy to spot.

Where does my recording go?

The sound is taken out of it in your browser; a video’s picture never leaves your device. The sound is sent a minute at a time to our server, where Cloudflare runs Whisper to write it out. Neither the sound nor the text is kept, and the words go to our reader only to add the pinyin.

Can it transcribe a YouTube video?

Our browser extension writes captions for YouTube videos that have none, with pinyin and every word explained as the video plays.

More free tools

• Chinese speech to text: speak and see it written
• Chinese image to text: the Chinese in a picture
• Chinese text to speech: hear text read aloud
• Pinyin converter
• Dictation practice