Soniox Pro SDK

The most powerful, flexible, and blazing-fast Python SDK for the Soniox Speech-to-Text API.

Production-ready, fully-typed, and optimised for performance.

Features

Complete API Coverage

• REST API (Files, Transcriptions, Models, Authentication)
• WebSocket API (Real-time transcription and translation)
• Synchronous and asynchronous interfaces
• Full type safety with Pydantic models

Performance Optimised

• Blazing fast with optional C extensions
• Connection pooling for HTTP and WebSocket
• Async I/O throughout
• Memory-efficient streaming

Advanced Capabilities

• 60+ languages supported
• Real-time translation (one-way and two-way)
• Speaker diarization (up to 15 speakers)
• Language identification
• Endpoint detection
• Custom context and vocabulary
• Word-level timestamps
• Confidence scores

Developer Experience

• Comprehensive documentation
• Full IDE autocomplete support
• 90%+ test coverage
• Type-checked with mypy
• Clean, Pythonic API

Installation

Basic Installation

# Using uv (recommended)
uv add soniox-pro-sdk

# Using pip
pip install soniox-pro-sdk

Optional Dependencies

# Async support
pip install soniox-pro-sdk[async]

# Performance optimisations (C extensions)
pip install soniox-pro-sdk[performance]

# Development tools
pip install soniox-pro-sdk[dev]

# Everything
pip install soniox-pro-sdk[all]

Quick Start

Async Transcription

from soniox import SonioxClient

# Initialise client
client = SonioxClient(api_key="your-api-key")

# Transcribe from URL
transcription = client.transcriptions.create(
    audio_url="https://example.com/audio.mp3",
    model="stt-async-v3",
    enable_speaker_diarization=True,
)

# Wait for completion
result = client.transcriptions.wait_for_completion(transcription.id)
print(result.transcript.text)

Real-time Transcription

from soniox import SonioxRealtimeClient

# Initialise real-time client
client = SonioxRealtimeClient(
    api_key="your-api-key",
    model="stt-rt-v3",
)

# Stream audio
with client.stream() as stream:
    with open("audio.mp3", "rb") as f:
        while chunk := f.read(4096):
            stream.send_audio(chunk)

    # Receive tokens
    for response in stream:
        for token in response.tokens:
            if token.is_final:
                print(token.text, end="")

Async/Await Support

import asyncio
from soniox import AsyncSonioxClient

async def transcribe():
    async with AsyncSonioxClient(api_key="your-api-key") as client:
        file = await client.files.upload("audio.mp3")
        transcription = await client.transcriptions.create(file_id=file.id)
        result = await client.transcriptions.wait_for_completion(transcription.id)
        return result.transcript.text

text = asyncio.run(transcribe())

Documentation

Full documentation available at: https://github.com/CodeWithBehnam/soniox-pro-sdk

Testing

# Run tests
uv run pytest

# With coverage
uv run pytest --cov=soniox

# Type check
uv run mypy src/soniox

# Lint
uv run ruff check src/soniox

License

MIT License - see LICENSE for details.

Links

• Documentation: https://github.com/CodeWithBehnam/soniox-pro-sdk
• Source Code: https://github.com/CodeWithBehnam/soniox-pro-sdk
• Soniox API: https://soniox.com/docs

Built for the developer community