> ## Documentation Index
> Fetch the complete documentation index at: https://tennda.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Realtime

> WebSocket real-time audio conversation

## Introduction

OpenAI-compatible Realtime API over **WebSocket** for low-latency voice interaction—not a standard HTTP POST endpoint.

## Authentication

Pass your Bearer token when opening the WebSocket (per OpenAI Realtime conventions—header or initial auth message).

## Connection example

```
wss://api.tennda.ai/v1/realtime?model=gpt-4o-realtime-preview
```

<ParamField query="model" type="string">
  Realtime model, e.g. `gpt-4o-realtime-preview`
</ParamField>

## Notes

* **101 Switching Protocols** indicates a successful upgrade
* Exchange audio/text using Realtime event messages after connect
* Suitable for voice assistants and live translation

See [OpenAI Realtime guide](https://platform.openai.com/docs/guides/realtime) for event schemas.