Skip to content

PyAI Hear · Speech to text

Every word matters.
Hear it clearly.

Bring the spoken conversation into your application. Transcribe phone audio, follow live speech and turn recordings into text your team can use.

Your audio. A readable transcript. One speech API.

Your audio → Your transcriptInteractive demo
Try your audio. Take the code.

Upload a short audio clip and read the transcript. The code follows your settings.

Try a provided English recording

Recorded synthetic speech from Cast, supplied as input. Choose Transcribe to get a new Hear result.

WAV, MP3, M4A, FLAC, or OGG · Up to 8 MB. Audio is sent to PyAI only when you choose Transcribe.

No signup

Live demo: up to 8 requests per day.

import { readFile } from "node:fs/promises";

const filePath = "audio.wav";
const form = new FormData();
form.append("file", new Blob([await readFile(filePath)]), filePath);
form.append("model", "pyai-hear");
form.append("response_format", "json");

const response = await fetch("https://api.pyai.com/v1/audio/transcriptions", {
  method: "POST",
  headers: { Authorization: `Bearer ${process.env.PYAI_API_KEY}` },
  body: form,
});
if (!response.ok) throw new Error(await response.text());
console.log(await response.json());

Set PYAI_API_KEY in your server environment. Put the selected file beside your script, or update its path.

Copy the code to use these settings. You can also open Playground and choose your audio there.

Open Playground

A real Hear API request, started only when you choose Transcribe.

Built for narrowband phone audioLive streaming and recorded audioOpenAI-compatible API

Hear + Speak · Live demo

Ask a question. Hear the answer.

Record a question about PyAI, see Hear transcribe it, and listen to Speak read the reply. You can also type. The product answers are scripted; the transcription and speech use real PyAI APIs.

Three free questions, no signup required. Daily demo limits apply. Ready to build? Sign up from the demo or use your own key with the integration guide.

Available on ClawHub

Use PyAI Hear in OpenClaw.

Install the skills you need, connect your PyAI account through the official CLI, and ask your assistant to put them to work.

Hear · v1.0.0

Turn a voice note into useful text.

Transcribe a supplied recording, interview or voicemail and return the words you can search, quote or work with.

Get Hear on ClawHub

A PyAI account and transcription permissions. Audio processing consumes usage.

Create an account for Hear

From sound to something useful

More than a recording.
A conversation you can use.

One API for a short voicemail, a live stream or an archive of conversations. Choose the path that fits the job.

01

Give every recording a next step.

Upload a clip for an immediate transcript, or send a longer recording to an asynchronous job. Use async diarization for speaker-labelled segments and timestamps.

Read the transcription guide
02

Read the conversation as it happens.

Stream audio over a WebSocket and receive partials as the caller speaks. Partials are revisable; final events let your application commit the utterance.

Explore streaming transcription
03

Make the final words more useful.

On English finals, opt in to punctuation, numeral formatting, filler removal and vocabulary. Live partials stay fast; cleanup applies to the final text.

Explore Hear configuration

Made for the pace of conversation

Catch the words.
Keep the conversation.

Hear is tuned for telephone audio. Streaming partials help your interface respond as people speak, while final events give you a clear point to commit the text.

Explore supported languages
Streaming transcription

~200ms

Time to first partial

Measured in-region. Partial text can change as more speech arrives; a first partial is different from a completed transcript.

WAVMP3M4AFLAC

Fits the way you build

Keep your application.
Change its voice layer.

Use the familiar OpenAI transcription request shape for recorded audio. For live speech, connect through Hear’s native streaming protocol.

ONE REQUEST, A USEFUL RESULT

Your audioHearYour text
  • Pick a language or let Hear detect it
  • Choose a response format for your workflow
  • Use the transcript in your application
Configure the example above
Make it part ofContact centersVoicemailCall searchLive captionsVoice applications

Speaker diarization

Know the words.
See who said them.

Use Hear async jobs for speaker-labelled segments on mono recordings, or preserve separate participants in dual-channel audio.

Speaker diarization

Know the words.
See who said them.

Use Hear async jobs for speaker-labelled segments on mono recordings, or preserve separate participants in dual-channel audio.

A little more detail

Good questions.
Clear answers.

Sandbox keys never bill; live usage requires available credit. See current rates ↗

What is Hear?

Hear is PyAI's speech-to-text API. It transcribes 8 kHz telephone audio in batch or realtime, speaks the OpenAI transcription contract, and can clean English finals (digits, punctuation, fillers, vocabulary) without touching the live partials.

Does PyAI support speaker diarization?

Yes. Hear async transcription jobs support speaker diarization for mono recordings with diarize: true. For recordings with one participant on each stereo channel, use channel: true instead. Results include speaker-labelled segments and timestamps. Labels identify speakers within a recording, not verified identities or agent/customer roles. Check that the returned segments contain speaker labels before relying on them.

How much does Hear STT cost?

You pay only for audio transcribed, billed per minute - no seats, no platform fee, no minimums. Check our pricing page for the latest rates.

How do I migrate from OpenAI Whisper?

Change two lines: point your existing OpenAI SDK at https://api.pyai.com/v1 and pass your PyAI key. The request and response shapes match, so the rest of your code is unchanged.

Does Hear stream?

Yes. Realtime STT runs over a WebSocket (GET /v1/audio/transcriptions/stream?protocol=pyai-hear-v1), returning partial results as the caller speaks and a final result per utterance.

When is Hear the right fit?

Hear is built for phone calls. If you transcribe narrowband 8 kHz call audio - IVR, contact-center, voicemail, real-time agents - Hear is tuned for it where studio-trained STT degrades. For long recordings you don't need instantly, the async batch tier halves the cost.

Is it accurate on bad phone lines?

Hear is trained on real narrowband call audio rather than clean studio takes, so the accuracy holds up on the 8 kHz lines and codecs that real telephony actually uses.

The next thing you build

Give your audio
a next chapter.