Official BytePlus partnerBuilt on BytePlus, ByteDance's official AI platform

Text to speech · Mandarin

Mandarin text to speech with 8 AI voices

Paste a Mandarin script, pick one of 8 preset voices (Cora (female), Suki (female), Sophia (female), Qiang (male), Lan (female), Ming (male), Bo (male) and Wen (female)) and get natural speech in seconds. Set the speed, pitch and volume, write a two-speaker dialogue and download the clip.

  • 8 Mandarin voices
  • Up to 5,000 characters per clip
  • 1 credit per 1,000 characters

The 8 Mandarin voices

Every voice under its real studio name, with its gender and use-case tags from the catalogue. Open it in Text to Speech and hear it read your own script.

Accents and regional variants

The 8 preset voices are listed in the catalogue as Mandarin, without a regional label. Cantonese is available in Audio Generation, which speaks 24 languages from a written prompt: pick the variant in the language menu and the model speaks it.

In Audio Generation the voice comes from the prompt, not from a list, so it can change between runs. For a fixed, reusable voice, use the preset voices on this page.

Open Audio Generation

What Mandarin text to speech is good for

  • Lively social posts and short video commentary
  • Audiobooks and storytelling with a deep narrator
  • Character roles and late-night podcast style shows
  • Warm customer messages and brand promos

Tips for Mandarin scripts

  • Write the script in Chinese characters, not pinyin.
  • Use Chinese punctuation (,。!?) to set the pauses and the tone of each sentence.
  • Write numbers, dates and amounts the way they should be read when the reading matters.

Two-speaker dialogue

In "Two speakers" mode, cast Cora as Speaker A and Suki as Speaker B and write up to 12 turns. Each turn is generated in its voice and the conversation plays back in order. You can also pair a Mandarin voice with a voice in another language.

More Mandarin audio tools

Speech to Text

Transcribe Mandarin recordings into editable text, with auto-detect or the language picked.

Live Interpretation

Translate a voice clip into Mandarin and hear the translation read by a built-in voice.

Audio Generation

Describe a scene in a prompt and the model performs it in Mandarin, with cued lines and timing.

Voice Replication

Clone your own voice with your consent recorded and use it in Text to Speech next to the preset voices.

Pricing

Text to Speech costs 1 credit per 1,000 characters, with a minimum of 1 credit per clip: an 800-character Mandarin script costs 1 credit and a full 5,000-character script costs 5 credits. The studio shows the cost before you generate, and failed clips are refunded automatically.

Mandarin text to speech questions

How many Mandarin voices are there?

8: Cora (female), Suki (female), Sophia (female), Qiang (male), Lan (female), Ming (male), Bo (male) and Wen (female). Across all languages, Text to Speech has 45 voices in 14 languages.

Are there regional Mandarin accents?

The preset voices are listed as Mandarin, without a regional label. Cantonese is available in Audio Generation, steered by the prompt, not as a preset voice.

How much does it cost?

Text to Speech costs 1 credit per 1,000 characters, with a minimum of 1 credit per clip: an 800-character Mandarin script costs 1 credit and a full 5,000-character script costs 5 credits. The studio shows the cost before you generate, and failed clips are refunded automatically.

Do the Mandarin voices support emotions?

The Mandarin voices have no emotion styles. Set the speed between 0.5x and 2x, the pitch and the volume to shape the read.

Can I transcribe or translate Mandarin audio?

Yes, Speech to Text transcribes Mandarin. Live Interpretation translates voice clips into Mandarin.

Which platform does the audio studio run on?

Built on BytePlus, ByteDance's official AI platform. KreatorFlow is an official BytePlus partner; the audio studio and its credit prices are KreatorFlow's own.

Hear your Mandarin script

Pick one of 8 voices, paste your text and generate. 1 credit per 1,000 characters.

Text to speech in other languages