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

# Openai

> JavaScript API reference for the openai realtime models plugin.

## OpenAIRealtime

OpenAI Realtime speech-to-speech (S2S) model.

### Constructor

```typescript theme={null}
new OpenAIRealtime(opts: { apiKey?: string; config?: OpenAIRealtimeConfig | null; [key: string]: any })
```

<ParamField path="apiKey" type="string" />

<ParamField path="config" type="OpenAIRealtimeConfig | null" />

### cleanup

```typescript theme={null}
cleanup(): Promise<void>
```

### close

```typescript theme={null}
close(): Promise<void>
```

### getRuntimeConfig

```typescript theme={null}
getRuntimeConfig(): Record<string, any>
```

<ResponseField name="returns" type="Record<string, any>" />

***

## SUPPORTED\_VOICES

Voice ids supported by the OpenAI Realtime model.

```typescript theme={null}
const SUPPORTED_VOICES = [ 'alloy', 'ash', 'ballad', 'breeze', 'cedar', 'cinnamon', 'coral', 'echo', 'ember', 'juniper', 'marin', 'sage', 'shimmer', 'verse', ] as const
```

***

## TurnDetectionConfig

Server-side voice-activity turn-detection settings. Pass `null` to disable.

### Fields

<ParamField path="type" type="string">
  Detection mode. Default `'server_vad'`.
</ParamField>

<ParamField path="threshold" type="number">
  Speech-probability threshold in \[0, 1]. Default `0.5`.
</ParamField>

<ParamField path="prefixPaddingMs" type="number">
  Audio padding (ms) prepended before detected speech. Default `300`.
</ParamField>

<ParamField path="silenceDurationMs" type="number">
  Trailing silence (ms) that ends a turn. Default `200`.
</ParamField>

<ParamField path="createResponse" type="boolean">
  Auto-create a model response when a turn ends. Default `true`.
</ParamField>

<ParamField path="interruptResponse" type="boolean">
  Allow user speech to interrupt the model's response. Default `true`.
</ParamField>

***

## InputAudioTranscriptionConfig

Input-audio transcription settings. Pass `null` to disable transcription.

### Fields

<ParamField path="model" type="string">
  Transcription model id. Default `'gpt-4o-mini-transcribe'`.
</ParamField>

***

## OpenAIRealtimeConfig

Configuration for OpenAIRealtime.

### Fields

<ParamField path="model" type="string">
  Realtime model id. Default `'gpt-4o-realtime-preview'`.
</ParamField>

<ParamField path="voice" type="string">
  Voice id (see SUPPORTED\_VOICES). Default `'alloy'`.
</ParamField>

<ParamField path="modalities" type="string[]">
  Response modalities. Default `['text', 'audio']`.
</ParamField>

<ParamField path="temperature" type="number">
  Sampling temperature. Default `0.8`.
</ParamField>

<ParamField path="maxResponseOutputTokens" type="number | string | null">
  Max output tokens per response. Default the string `'inf'` (unlimited);
  `null` omits the param.
</ParamField>

<ParamField path="turnDetection" type="TurnDetectionConfig | null">
  Turn-detection settings, or `null` to disable. Defaults applied when omitted.
</ParamField>

<ParamField path="inputAudioTranscription" type="InputAudioTranscriptionConfig | null">
  Input-audio transcription settings, or `null` to disable. Default model applied when omitted.
</ParamField>

<ParamField path="toolChoice" type="string">
  Tool-selection strategy. Default `'auto'`.
</ParamField>