wehub-resource-sync
64 lines
1.8 KiB
Markdown
64 lines
1.8 KiB
Markdown
# Soniox plugin for LiveKit Agents
|
|
|
|
Support for Soniox [Speech-to-Text](https://soniox.com/docs/stt) and [Text-to-Speech](https://soniox.com/docs/tts) APIs, using WebSocket streaming interfaces.
|
|
|
|
See [STT documentation](https://docs.livekit.io/agents/models/stt/soniox/) and [TTS documentation](https://docs.livekit.io/agents/models/tts/soniox/) for more information.
|
|
|
|
## Installation
|
|
|
|
```bash
|
|
pip install livekit-plugins-soniox
|
|
```
|
|
|
|
## Pre-requisites
|
|
|
|
The Soniox plugin requires an API key to authenticate. You can get your Soniox API key [here](https://console.soniox.com/).
|
|
|
|
Set API key in your `.env` file:
|
|
|
|
```
|
|
SONIOX_API_KEY=<your_soniox_api_key>
|
|
```
|
|
|
|
## Usage
|
|
|
|
### Speech-to-Text (STT)
|
|
|
|
Use Soniox STT in an `AgentSession` or as a standalone transcription service:
|
|
|
|
```python
|
|
from livekit.plugins import soniox
|
|
|
|
session = AgentSession(
|
|
stt=soniox.STT(),
|
|
# ... llm, tts, etc.
|
|
)
|
|
```
|
|
|
|
### Text-to-Speech (TTS)
|
|
|
|
Use Soniox TTS for real-time speech synthesis:
|
|
|
|
```python
|
|
from livekit.plugins import soniox
|
|
|
|
session = AgentSession(
|
|
tts=soniox.TTS(
|
|
language="en",
|
|
voice="Maya",
|
|
speed=1.0, # 0.7-1.3
|
|
),
|
|
# ... stt, llm, etc.
|
|
)
|
|
```
|
|
|
|
The TTS supports real-time streaming from LLM - text chunks are tokenized and sent to Soniox as words are formed, enabling low-latency speech synthesis.
|
|
|
|
## More information and reference
|
|
|
|
Explore integration details and find comprehensive examples:
|
|
- [Soniox STT LiveKit integration guide](https://soniox.com/docs/integrations/livekit)
|
|
- [Soniox API reference](https://soniox.com/docs/api-reference)
|
|
- [Soniox STT languages](https://soniox.com/docs/stt/concepts/supported-languages) and [Soniox TTS languages](https://soniox.com/docs/tts/concepts/languages)
|
|
- [Soniox TTS voices](https://soniox.com/docs/tts/concepts/voices)
|