Mandarin speech to text, built on PyAI Hear
Mandarin speech to text turns spoken Mandarin into simplified Chinese your software can store and search. PyAI Hear is the engine that reads it, the same one behind live phone agents on PyAI. Chinese is written with no spaces between words, so Hear has to segment the characters as it recognizes them, and it uses context to pick the right character where tones alone would be ambiguous. Hear reads clean recordings and rough phone lines, and returns plain text, JSON or timestamped segments that stay in your own flow. Use it to transcribe Mandarin audio to text from a saved file, or as Mandarin voice to text for a clip you record right here on the page. As an audio to text generator, it suits one off files as well as steady pipelines. PyAI works as a free Mandarin speech to text generator, so a saved file or a quick recording turns into text you can edit, search and reuse.
Key takeaways
- Price: $0.001/min, billed by the whole minute.
- Simplified output: Hear returns simplified Chinese characters with no spaces between words, the way the text is actually read, so a transcript drops straight into a Chinese interface without any post cleaning.
- Context over tone: where two words sound alike and only tone or context tells them apart, the engine uses the surrounding sentence to choose the right character.
- No lock in: no platform fee, so you pay for the Mandarin minutes you send and nothing else.
- Accuracy: word error rate is 1.59% on clean audio and 4.76% on 8 kHz phone audio, PyAI's own benchmark numbers measured on English and reported as industry parity, not a per language score.
- Drop in: a whisper-1 client reaches Hear by changing the base URL, first partial in about 200 ms, files up to 25 MB per call.
How Mandarin transcription works
You send Mandarin audio to one endpoint and get text back. The same engine runs as a file call, a live stream, or a background job.
Post a recorded Mandarin file to the endpoint, or stream PCM16 frames over a WebSocket for a live call or a meeting.
The engine segments the characters as it goes, uses context to pick between words that sound alike, marks each speaker, and firms up partial lines as the sentence arrives.
Filler is removed, dictation commands are honored, and numbers are written as digits so an amount or a date reads cleanly in a Chinese transcript.
Simplified Chinese in plain text, JSON, or verbose JSON with per segment timestamps and speaker labels, ready for notes, captions or search.
What Hear handles in Mandarin
Mandarin brings its own accents, numbers and names. Here is how Hear deals with them.
No spaces between words
Chinese is written without spaces, so Hear segments the stream of characters itself and returns text that reads naturally rather than one unbroken run you have to split later.
Tones and look alike words
Many Mandarin syllables share a sound and differ only by tone or by meaning. The engine reads the whole sentence to settle on the character a speaker meant.
Numbers and names
Amounts, dates and order numbers come out as digits, and up to 5 custom vocabulary terms teach it the product names and places your audio keeps repeating.
Phone and noisy audio
A lot of Mandarin audio is an 8 kHz call from a factory floor or a street. PyAI measures 4.76% word error rate on 8 kHz phone audio in English, so test your own Mandarin calls before launch.
Several speakers
A turn detector separates a host and two colleagues on a standup, so each voice returns as its own labeled segment instead of a single block of text.
Long recordings
A two hour webinar is chunked, transcribed in parallel and stitched back together, so a backlog of recordings clears at 92 to 247 times real time.
Accuracy and speed
Hear is at industry parity for accuracy, held on the hard case of real phone audio. The figures below are PyAI's own measurements on English speech, not a per language claim.
What you can transcribe in Mandarin
The same Mandarin endpoint feeds notes, captions, search and analytics.
Support and sales calls
Log every Mandarin call as searchable simplified Chinese text, then feed it to quality review, coaching or a CRM note without a person typing it out.
Team meetings
Stream a live standup or planning call and read a running transcript with speaker labels, so tasks are captured in Chinese as they are agreed.
Video captions
Make Mandarin captions with per segment timestamps that drop into an SRT or VTT file for a course or a product video.
Voice notes to tasks
Someone speaks a quick Mandarin update and it becomes text in a task or ticket, filler removed, ready to act on.
Webinars and interviews
Transcribe long Mandarin audio as a background job and get timestamped simplified Chinese you can quote, clip or index.
Voice search in an app
Send a short Mandarin clip and act on the text, so a user can speak a query or a command in Chinese inside your product.
Formats, limits and output
| Input | Audio file up to 25 MB for a single file call, or a live PCM16 stream over WebSocket for real time. OpenAI compatible, so a whisper-1 client works unchanged. |
|---|---|
| Output | Plain text, JSON, or verbose JSON with per segment timestamps. Smart formatting, dictation and filler removal are applied on the way out. |
| Modes | Sync file call for short clips, live streaming for calls and meetings, and background jobs for long recordings that run 92 to 247 times real time. |
| Tuning | Set the language to Mandarin, add numerals handling, and load up to 5 custom vocabulary terms so names and product words land right. |
| Billing | $0.001/min of audio, rounded to the whole minute. An empty transcript costs nothing, and a busy engine returns a retry, not a charge. |
Call Mandarin speech to text from your own code
Hear speaks the OpenAI transcription format, so most code points at PyAI by changing the base URL. Full docs: PyAI SDKs.
SDK
From a server, the SDK posts a Mandarin file or opens a live stream and returns simplified Chinese in a few lines. The SDK speaks the whisper-1 shape, so transcription code you already run moves over by swapping the base URL.
MCP
With the MCP server, an agent can transcribe a Mandarin recording on its own inside a flow, for example turning a meeting recording into notes, without you wiring the API call by hand.
CLI
From the terminal, the CLI takes a Mandarin audio file or a whole folder and writes out simplified Chinese text or timestamped JSON, which suits batch clearing a backlog of recordings.
Where Hear fits
PyAI is voice infrastructure. Hear, Speak and Omni run as one stack in one region, so recognition, reasoning and reply share a session instead of handing audio between vendors. Hear is not a consumer dictation app, and PyAI does not resell another company's recognizer under a new name. Hear is the Mandarin transcription engine you call from your own product, on your own calls and files. The accuracy figures are published on PyAI benchmarks.
Mandarin speech to text FAQ
How accurate is Mandarin speech to text on Hear?
Hear reports 1.59% word error rate on clean audio and 4.76% on 8 kHz phone audio. Both figures are PyAI's own benchmark numbers measured on English and held at industry parity, so they describe the engine overall and are not a separate Mandarin figure.
Does Hear write Mandarin with correct word segmentation?
Yes. Chinese has no spaces between words, so Hear segments the stream of characters as it recognizes the audio. You get simplified Chinese that reads naturally and fits a Chinese interface, not one long run of characters to split yourself.
How does Hear handle tones and look alike words?
Many Mandarin syllables sound the same and differ only by tone or meaning. The engine reads the surrounding sentence to pick the character a speaker meant, so context, not sound alone, settles which word lands in the transcript. No Mandarin accuracy figure is published, so test your own clips.
How much does Mandarin transcription cost?
Hear bills $0.001/min of audio, rounded to the whole minute, and an empty transcript is free. A new account opens with $50 of credit and no credit card, which covers a lot of Mandarin test audio before you spend anything, and sandbox keys never bill.
Can my whisper-1 code transcribe Mandarin?
Yes. Hear is OpenAI compatible, so a client written for whisper-1 reaches it by pointing at the PyAI base URL. Your request shape stays the same and you get text, JSON or verbose JSON with timestamps back as before, so no rewrite is needed.
Prices checked November 2026. Hear figures from pyai.com/benchmarks.