speechmatics-rt

What it is and what it does

This is an async-first Python client library for Speechmatics' real-time speech-to-text API. It wraps WebSocket communication to the Speechmatics service, allowing you to send audio streams and receive transcription results as they arrive. The library handles connection lifecycle, authentication (via API key or JWT), and event dispatch for transcript updates.

The package supports two main use patterns: single-stream transcription for simple cases where you're processing one audio source, and multi-channel transcription for scenarios where you need to simultaneously process multiple audio inputs with channel identification and diarization. Both patterns use an event-driven architecture where you register handlers for transcript events and let the client invoke them as results arrive. Type hints throughout the codebase provide IDE support and runtime safety.

Use it for:

• Build a real-time call transcription service that processes multiple participant audio streams simultaneously and labels output by speaker.
• Create a live meeting assistant that transcribes audio as it streams and triggers downstream actions on transcript events.
• Implement a voice command interface that sends microphone input to Speechmatics and reacts to recognized commands.
• Develop a multi-language customer support dashboard that transcribes calls from multiple channels in real time.
• Integrate speech-to-text into a browser application using JWT tokens for secure, short-lived authentication.

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Async Python client for real-time speech-to-text transcription via the Speechmatics API, supporting both single-stream and multi-channel audio processing over WebSocket.

Yes. The package is actively maintained, has low install friction, carries no known vulnerabilities, and offers a clean async interface for a specific task (real-time Speechmatics transcription). Install it if you need to integrate Speechmatics' real-time API; skip it if you're using a different speech-to-text provider or don't need real-time streaming.

Install

speechmatics-rt on PyPI

pip install speechmatics-rt

uv add speechmatics-rt

poetry add speechmatics-rt

Installing speechmatics-rt

Before you install

Low friction install with only two runtime dependencies (websockets and typing-extensions). Package is actively maintained with recent commits and is in beta status, having been released since June 2025.

License in practice

MIT license permits commercial and private use with minimal restrictions, making it suitable for most projects.

Quickstart

pip install speechmatics-rt

import asyncio
from speechmatics.rt import AsyncClient, ServerMessageType

async def main():
    async with AsyncClient() as client:
        @client.on(ServerMessageType.ADD_TRANSCRIPT)
        def handle_transcript(msg):
            print(msg['metadata']['transcript'])
        with open('audio.wav', 'rb') as f:
            await client.transcribe(f)

asyncio.run(main())

Requires Python 3.9 or later; SPEECHMATICS_API_KEY environment variable or explicit authentication must be configured.

Verify before relying

• Actual latency and throughput characteristics for real-time transcription workloads.
• Supported audio formats, sample rates, and codec requirements.
• Rate limits and quota behavior under sustained multi-channel load.
• Retry and reconnection behavior for network interruptions.

Package facts

Evidence: speechmatics_rt-1.1.0-py3-none-any.whl

Keywords: speechmatics, speech-to-text, real-time, transcription, websocket

Tags

More Python Modules packages