@livekit/agents-plugin-openai 0.9.0 → 0.9.2

package/dist/models.d.cts

@@ -0,0 +1,15 @@
+export type ChatModels = 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4-vision-preview' | 'gpt-4-1106-vision-preview' | 'gpt-4' | 'gpt-4-0314' | 'gpt-4-0613' | 'gpt-4-32k' | 'gpt-4-32k-0314' | 'gpt-4-32k-0613' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-16k' | 'gpt-3.5-turbo-0301' | 'gpt-3.5-turbo-0613' | 'gpt-3.5-turbo-1106' | 'gpt-3.5-turbo-16k-0613';
+export type WhisperModels = 'whisper-1';
+export type TTSModels = 'tts-1' | 'tts-1-hd' | 'gpt-4o-mini-tts';
+export type TTSVoices = 'alloy' | 'ash' | 'ballad' | 'coral' | 'echo' | 'fable' | 'nova' | 'onyx' | 'sage' | 'shimmer' | 'verse';
+export type TelnyxChatModels = 'meta-llama/Meta-Llama-3.1-8B-Instruct' | 'meta-llama/Meta-Llama-3.1-70B-Instruct';
+export type CerebrasChatModels = 'llama3.1-8b' | 'llama3.1-70b';
+export type PerplexityChatModels = 'llama-3.1-sonar-small-128k-online' | 'llama-3.1-sonar-small-128k-chat' | 'llama-3.1-sonar-large-128k-online' | 'llama-3.1-sonar-large-128k-chat' | 'llama-3.1-8b-instruct' | 'llama-3.1-70b-instruct';
+export type GroqChatModels = 'llama-3.1-405b-reasoning' | 'llama-3.1-70b-versatile' | 'llama-3.1-8b-instant' | 'llama-3.3-70b-versatile' | 'llama3-groq-70b-8192-tool-use-preview' | 'llama3-groq-8b-8192-tool-use-preview' | 'llama-guard-3-8b' | 'llama3-70b-8192' | 'llama3-8b-8192' | 'mixtral-8x7b-32768' | 'gemma-7b-it' | 'gemma2-9b-it';
+export type GroqAudioModels = 'whisper-large-v3' | 'distil-whisper-large-v3-en' | 'whisper-large-v3-turbo';
+export type DeepSeekChatModels = 'deepseek-coder' | 'deepseek-chat';
+export type TogetherChatModels = 'garage-bAInd/Platypus2-70B-instruct' | 'google/gemma-2-27b-it' | 'google/gemma-2-9b-it' | 'google/gemma-2b-it' | 'google/gemma-7b-it' | 'lmsys/vicuna-13b-v1.5' | 'lmsys/vicuna-7b-v1.5' | 'meta-llama/Llama-2-13b-chat-hf' | 'meta-llama/Llama-2-70b-chat-hf' | 'meta-llama/Llama-2-7b-chat-hf' | 'meta-llama/Llama-3-70b-chat-hf' | 'meta-llama/Llama-3-8b-chat-hf' | 'meta-llama/Meta-Llama-3-70B-Instruct-Lite' | 'meta-llama/Meta-Llama-3-70B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3-8B-Instruct-Lite' | 'meta-llama/Meta-Llama-3-8B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo' | 'mistralai/Mistral-7B-Instruct-v0.1' | 'mistralai/Mistral-7B-Instruct-v0.2' | 'mistralai/Mistral-7B-Instruct-v0.3' | 'mistralai/Mixtral-8x22B-Instruct-v0.1' | 'mistralai/Mixtral-8x7B-Instruct-v0.1' | 'openchat/openchat-3.5-1210' | 'snorkelai/Snorkel-Mistral-PairRM-DPO' | 'teknium/OpenHermes-2-Mistral-7B' | 'teknium/OpenHermes-2p5-Mistral-7B' | 'togethercomputer/Llama-2-7B-32K-Instruct' | 'togethercomputer/RedPajama-INCITE-7B-Chat' | 'togethercomputer/RedPajama-INCITE-Chat-3B-v1' | 'togethercomputer/StripedHyena-Nous-7B' | 'togethercomputer/alpaca-7b' | 'upstage/SOLAR-10.7B-Instruct-v1.0' | 'zero-one-ai/Yi-34B-Chat';
+export type OctoChatModels = 'meta-llama-3-70b-instruct' | 'meta-llama-3.1-405b-instruct' | 'meta-llama-3.1-70b-instruct' | 'meta-llama-3.1-8b-instruct' | 'mistral-7b-instruct' | 'mixtral-8x7b-instruct' | 'wizardlm-2-8x22bllamaguard-2-7b';
+export type XAIChatModels = 'grok-2' | 'grok-2-mini' | 'grok-2-mini-public' | 'grok-2-public';
+export type MetaChatModels = 'Llama-4-Scout-17B-16E-Instruct-FP8' | 'Llama-4-Maverick-17B-128E-Instruct-FP8' | 'Llama-3.3-70B-Instruct' | 'Llama-3.3-8B-Instruct';
+//# sourceMappingURL=models.d.ts.map

package/dist/models.d.ts

@@ -1,7 +1,7 @@
 export type ChatModels = 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-mini' | 'gpt-4o-mini-2024-07-18' | 'gpt-4-turbo' | 'gpt-4-turbo-2024-04-09' | 'gpt-4-turbo-preview' | 'gpt-4-0125-preview' | 'gpt-4-1106-preview' | 'gpt-4-vision-preview' | 'gpt-4-1106-vision-preview' | 'gpt-4' | 'gpt-4-0314' | 'gpt-4-0613' | 'gpt-4-32k' | 'gpt-4-32k-0314' | 'gpt-4-32k-0613' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-16k' | 'gpt-3.5-turbo-0301' | 'gpt-3.5-turbo-0613' | 'gpt-3.5-turbo-1106' | 'gpt-3.5-turbo-16k-0613';
 export type WhisperModels = 'whisper-1';
-export type TTSModels = 'tts-1' | 'tts-1-hd';
-export type TTSVoices = 'alloy' | 'echo' | 'fable' | 'onyx' | 'nova' | 'shimmer';
+export type TTSModels = 'tts-1' | 'tts-1-hd' | 'gpt-4o-mini-tts';
+export type TTSVoices = 'alloy' | 'ash' | 'ballad' | 'coral' | 'echo' | 'fable' | 'nova' | 'onyx' | 'sage' | 'shimmer' | 'verse';
 export type TelnyxChatModels = 'meta-llama/Meta-Llama-3.1-8B-Instruct' | 'meta-llama/Meta-Llama-3.1-70B-Instruct';
 export type CerebrasChatModels = 'llama3.1-8b' | 'llama3.1-70b';
 export type PerplexityChatModels = 'llama-3.1-sonar-small-128k-online' | 'llama-3.1-sonar-small-128k-chat' | 'llama-3.1-sonar-large-128k-online' | 'llama-3.1-sonar-large-128k-chat' | 'llama-3.1-8b-instruct' | 'llama-3.1-70b-instruct';
@@ -11,4 +11,5 @@ export type DeepSeekChatModels = 'deepseek-coder' | 'deepseek-chat';
 export type TogetherChatModels = 'garage-bAInd/Platypus2-70B-instruct' | 'google/gemma-2-27b-it' | 'google/gemma-2-9b-it' | 'google/gemma-2b-it' | 'google/gemma-7b-it' | 'lmsys/vicuna-13b-v1.5' | 'lmsys/vicuna-7b-v1.5' | 'meta-llama/Llama-2-13b-chat-hf' | 'meta-llama/Llama-2-70b-chat-hf' | 'meta-llama/Llama-2-7b-chat-hf' | 'meta-llama/Llama-3-70b-chat-hf' | 'meta-llama/Llama-3-8b-chat-hf' | 'meta-llama/Meta-Llama-3-70B-Instruct-Lite' | 'meta-llama/Meta-Llama-3-70B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3-8B-Instruct-Lite' | 'meta-llama/Meta-Llama-3-8B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo' | 'meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo' | 'mistralai/Mistral-7B-Instruct-v0.1' | 'mistralai/Mistral-7B-Instruct-v0.2' | 'mistralai/Mistral-7B-Instruct-v0.3' | 'mistralai/Mixtral-8x22B-Instruct-v0.1' | 'mistralai/Mixtral-8x7B-Instruct-v0.1' | 'openchat/openchat-3.5-1210' | 'snorkelai/Snorkel-Mistral-PairRM-DPO' | 'teknium/OpenHermes-2-Mistral-7B' | 'teknium/OpenHermes-2p5-Mistral-7B' | 'togethercomputer/Llama-2-7B-32K-Instruct' | 'togethercomputer/RedPajama-INCITE-7B-Chat' | 'togethercomputer/RedPajama-INCITE-Chat-3B-v1' | 'togethercomputer/StripedHyena-Nous-7B' | 'togethercomputer/alpaca-7b' | 'upstage/SOLAR-10.7B-Instruct-v1.0' | 'zero-one-ai/Yi-34B-Chat';
 export type OctoChatModels = 'meta-llama-3-70b-instruct' | 'meta-llama-3.1-405b-instruct' | 'meta-llama-3.1-70b-instruct' | 'meta-llama-3.1-8b-instruct' | 'mistral-7b-instruct' | 'mixtral-8x7b-instruct' | 'wizardlm-2-8x22bllamaguard-2-7b';
 export type XAIChatModels = 'grok-2' | 'grok-2-mini' | 'grok-2-mini-public' | 'grok-2-public';
+export type MetaChatModels = 'Llama-4-Scout-17B-16E-Instruct-FP8' | 'Llama-4-Maverick-17B-128E-Instruct-FP8' | 'Llama-3.3-70B-Instruct' | 'Llama-3.3-8B-Instruct';
 //# sourceMappingURL=models.d.ts.map

package/dist/models.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"models.d.ts","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAIA,MAAM,MAAM,UAAU,GAClB,QAAQ,GACR,mBAAmB,GACnB,aAAa,GACb,wBAAwB,GACxB,aAAa,GACb,wBAAwB,GACxB,qBAAqB,GACrB,oBAAoB,GACpB,oBAAoB,GACpB,sBAAsB,GACtB,2BAA2B,GAC3B,OAAO,GACP,YAAY,GACZ,YAAY,GACZ,WAAW,GACX,gBAAgB,GAChB,gBAAgB,GAChB,eAAe,GACf,mBAAmB,GACnB,oBAAoB,GACpB,oBAAoB,GACpB,oBAAoB,GACpB,wBAAwB,CAAC;AAE7B,MAAM,MAAM,aAAa,GAAG,WAAW,CAAC;AAExC,MAAM,MAAM,SAAS,GAAG,OAAO,GAAG,UAAU,CAAC;AAE7C,MAAM,MAAM,SAAS,GAAG,OAAO,GAAG,MAAM,GAAG,OAAO,GAAG,MAAM,GAAG,MAAM,GAAG,SAAS,CAAC;AAIjF,MAAM,MAAM,gBAAgB,GACxB,uCAAuC,GACvC,wCAAwC,CAAC;AAE7C,MAAM,MAAM,kBAAkB,GAAG,aAAa,GAAG,cAAc,CAAC;AAEhE,MAAM,MAAM,oBAAoB,GAC5B,mCAAmC,GACnC,iCAAiC,GACjC,mCAAmC,GACnC,iCAAiC,GACjC,uBAAuB,GACvB,wBAAwB,CAAC;AAE7B,MAAM,MAAM,cAAc,GACtB,0BAA0B,GAC1B,yBAAyB,GACzB,sBAAsB,GACtB,yBAAyB,GACzB,uCAAuC,GACvC,sCAAsC,GACtC,kBAAkB,GAClB,iBAAiB,GACjB,gBAAgB,GAChB,oBAAoB,GACpB,aAAa,GACb,cAAc,CAAC;AAEnB,MAAM,MAAM,eAAe,GACvB,kBAAkB,GAClB,4BAA4B,GAC5B,wBAAwB,CAAC;AAE7B,MAAM,MAAM,kBAAkB,GAAG,gBAAgB,GAAG,eAAe,CAAC;AAEpE,MAAM,MAAM,kBAAkB,GAC1B,qCAAqC,GACrC,uBAAuB,GACvB,sBAAsB,GACtB,oBAAoB,GACpB,oBAAoB,GACpB,uBAAuB,GACvB,sBAAsB,GACtB,gCAAgC,GAChC,gCAAgC,GAChC,+BAA+B,GAC/B,gCAAgC,GAChC,+BAA+B,GAC/B,2CAA2C,GAC3C,4CAA4C,GAC5C,0CAA0C,GAC1C,2CAA2C,GAC3C,+CAA+C,GAC/C,8CAA8C,GAC9C,6CAA6C,GAC7C,oCAAoC,GACpC,oCAAoC,GACpC,oCAAoC,GACpC,uCAAuC,GACvC,sCAAsC,GACtC,4BAA4B,GAC5B,sCAAsC,GACtC,iCAAiC,GACjC,mCAAmC,GACnC,0CAA0C,GAC1C,2CAA2C,GAC3C,8CAA8C,GAC9C,uCAAuC,GACvC,4BAA4B,GAC5B,mCAAmC,GACnC,yBAAyB,CAAC;AAE9B,MAAM,MAAM,cAAc,GACtB,2BAA2B,GAC3B,8BAA8B,GAC9B,6BAA6B,GAC7B,4BAA4B,GAC5B,qBAAqB,GACrB,uBAAuB,GACvB,iCAAiC,CAAC;AAEtC,MAAM,MAAM,aAAa,GAAG,QAAQ,GAAG,aAAa,GAAG,oBAAoB,GAAG,eAAe,CAAC"}
+{"version":3,"file":"models.d.ts","sourceRoot":"","sources":["../src/models.ts"],"names":[],"mappings":"AAIA,MAAM,MAAM,UAAU,GAClB,QAAQ,GACR,mBAAmB,GACnB,aAAa,GACb,wBAAwB,GACxB,aAAa,GACb,wBAAwB,GACxB,qBAAqB,GACrB,oBAAoB,GACpB,oBAAoB,GACpB,sBAAsB,GACtB,2BAA2B,GAC3B,OAAO,GACP,YAAY,GACZ,YAAY,GACZ,WAAW,GACX,gBAAgB,GAChB,gBAAgB,GAChB,eAAe,GACf,mBAAmB,GACnB,oBAAoB,GACpB,oBAAoB,GACpB,oBAAoB,GACpB,wBAAwB,CAAC;AAE7B,MAAM,MAAM,aAAa,GAAG,WAAW,CAAC;AAExC,MAAM,MAAM,SAAS,GAAG,OAAO,GAAG,UAAU,GAAG,iBAAiB,CAAC;AAEjE,MAAM,MAAM,SAAS,GACjB,OAAO,GACP,KAAK,GACL,QAAQ,GACR,OAAO,GACP,MAAM,GACN,OAAO,GACP,MAAM,GACN,MAAM,GACN,MAAM,GACN,SAAS,GACT,OAAO,CAAC;AAIZ,MAAM,MAAM,gBAAgB,GACxB,uCAAuC,GACvC,wCAAwC,CAAC;AAE7C,MAAM,MAAM,kBAAkB,GAAG,aAAa,GAAG,cAAc,CAAC;AAEhE,MAAM,MAAM,oBAAoB,GAC5B,mCAAmC,GACnC,iCAAiC,GACjC,mCAAmC,GACnC,iCAAiC,GACjC,uBAAuB,GACvB,wBAAwB,CAAC;AAE7B,MAAM,MAAM,cAAc,GACtB,0BAA0B,GAC1B,yBAAyB,GACzB,sBAAsB,GACtB,yBAAyB,GACzB,uCAAuC,GACvC,sCAAsC,GACtC,kBAAkB,GAClB,iBAAiB,GACjB,gBAAgB,GAChB,oBAAoB,GACpB,aAAa,GACb,cAAc,CAAC;AAEnB,MAAM,MAAM,eAAe,GACvB,kBAAkB,GAClB,4BAA4B,GAC5B,wBAAwB,CAAC;AAE7B,MAAM,MAAM,kBAAkB,GAAG,gBAAgB,GAAG,eAAe,CAAC;AAEpE,MAAM,MAAM,kBAAkB,GAC1B,qCAAqC,GACrC,uBAAuB,GACvB,sBAAsB,GACtB,oBAAoB,GACpB,oBAAoB,GACpB,uBAAuB,GACvB,sBAAsB,GACtB,gCAAgC,GAChC,gCAAgC,GAChC,+BAA+B,GAC/B,gCAAgC,GAChC,+BAA+B,GAC/B,2CAA2C,GAC3C,4CAA4C,GAC5C,0CAA0C,GAC1C,2CAA2C,GAC3C,+CAA+C,GAC/C,8CAA8C,GAC9C,6CAA6C,GAC7C,oCAAoC,GACpC,oCAAoC,GACpC,oCAAoC,GACpC,uCAAuC,GACvC,sCAAsC,GACtC,4BAA4B,GAC5B,sCAAsC,GACtC,iCAAiC,GACjC,mCAAmC,GACnC,0CAA0C,GAC1C,2CAA2C,GAC3C,8CAA8C,GAC9C,uCAAuC,GACvC,4BAA4B,GAC5B,mCAAmC,GACnC,yBAAyB,CAAC;AAE9B,MAAM,MAAM,cAAc,GACtB,2BAA2B,GAC3B,8BAA8B,GAC9B,6BAA6B,GAC7B,4BAA4B,GAC5B,qBAAqB,GACrB,uBAAuB,GACvB,iCAAiC,CAAC;AAEtC,MAAM,MAAM,aAAa,GAAG,QAAQ,GAAG,aAAa,GAAG,oBAAoB,GAAG,eAAe,CAAC;AAE9F,MAAM,MAAM,cAAc,GACtB,oCAAoC,GACpC,wCAAwC,GACxC,wBAAwB,GACxB,uBAAuB,CAAC"}

package/dist/realtime/api_proto.d.cts

@@ -0,0 +1,413 @@
+export declare const SAMPLE_RATE = 24000;
+export declare const NUM_CHANNELS = 1;
+export declare const IN_FRAME_SIZE = 2400;
+export declare const OUT_FRAME_SIZE = 1200;
+export declare const BASE_URL = "wss://api.openai.com/v1";
+export type Model = 'gpt-4o-realtime-preview-2024-10-01' | string;
+export type Voice = 'alloy' | 'shimmer' | 'echo' | 'ash' | 'ballad' | 'coral' | 'sage' | 'verse' | string;
+export type AudioFormat = 'pcm16';
+export type Role = 'system' | 'assistant' | 'user' | 'tool';
+export type GenerationFinishedReason = 'stop' | 'max_tokens' | 'content_filter' | 'interrupt';
+export type InputTranscriptionModel = 'whisper-1' | string;
+export type Modality = 'text' | 'audio';
+export type ToolChoice = 'auto' | 'none' | 'required' | string;
+export type State = 'initializing' | 'listening' | 'thinking' | 'speaking' | string;
+export type ResponseStatus = 'in_progress' | 'completed' | 'incomplete' | 'cancelled' | 'failed' | string;
+export type ClientEventType = 'session.update' | 'input_audio_buffer.append' | 'input_audio_buffer.commit' | 'input_audio_buffer.clear' | 'conversation.item.create' | 'conversation.item.truncate' | 'conversation.item.delete' | 'response.create' | 'response.cancel';
+export type ServerEventType = 'error' | 'session.created' | 'session.updated' | 'conversation.created' | 'input_audio_buffer.committed' | 'input_audio_buffer.cleared' | 'input_audio_buffer.speech_started' | 'input_audio_buffer.speech_stopped' | 'conversation.item.created' | 'conversation.item.input_audio_transcription.completed' | 'conversation.item.input_audio_transcription.failed' | 'conversation.item.truncated' | 'conversation.item.deleted' | 'response.created' | 'response.done' | 'response.output_item.added' | 'response.output_item.done' | 'response.content_part.added' | 'response.content_part.done' | 'response.text.delta' | 'response.text.done' | 'response.audio_transcript.delta' | 'response.audio_transcript.done' | 'response.audio.delta' | 'response.audio.done' | 'response.function_call_arguments.delta' | 'response.function_call_arguments.done' | 'rate_limits.updated';
+export type AudioBase64Bytes = string;
+export interface Tool {
+    type: 'function';
+    name: string;
+    description?: string;
+    parameters: {
+        type: 'object';
+        properties: {
+            [prop: string]: {
+                [prop: string]: any;
+            };
+        };
+        required: string[];
+    };
+}
+export type TurnDetectionType = {
+    type: 'server_vad';
+    threshold?: number;
+    prefix_padding_ms?: number;
+    silence_duration_ms?: number;
+};
+export type InputAudioTranscription = {
+    model: InputTranscriptionModel;
+};
+export interface InputTextContent {
+    type: 'input_text';
+    text: string;
+}
+export interface InputAudioContent {
+    type: 'input_audio';
+    audio: AudioBase64Bytes;
+}
+export interface TextContent {
+    type: 'text';
+    text: string;
+}
+export interface AudioContent {
+    type: 'audio';
+    audio: AudioBase64Bytes;
+    transcript: string;
+}
+export type Content = InputTextContent | InputAudioContent | TextContent | AudioContent;
+export type ContentPart = {
+    type: 'text' | 'audio';
+    audio?: AudioBase64Bytes;
+    transcript?: string;
+};
+export interface BaseItem {
+    id: string;
+    object: 'realtime.item';
+    type: string;
+}
+export interface SystemItem extends BaseItem {
+    type: 'message';
+    role: 'system';
+    content: InputTextContent;
+}
+export interface UserItem extends BaseItem {
+    type: 'message';
+    role: 'user';
+    content: (InputTextContent | InputAudioContent)[];
+}
+export interface AssistantItem extends BaseItem {
+    type: 'message';
+    role: 'assistant';
+    content: (TextContent | AudioContent)[];
+}
+export interface FunctionCallItem extends BaseItem {
+    type: 'function_call';
+    call_id: string;
+    name: string;
+    arguments: string;
+}
+export interface FunctionCallOutputItem extends BaseItem {
+    type: 'function_call_output';
+    call_id: string;
+    output: string;
+}
+export type ItemResource = SystemItem | UserItem | AssistantItem | FunctionCallItem | FunctionCallOutputItem;
+export interface SessionResource {
+    id: string;
+    object: 'realtime.session';
+    model: string;
+    modalities: ['text', 'audio'] | ['text'];
+    instructions: string;
+    voice: Voice;
+    input_audio_format: AudioFormat;
+    output_audio_format: AudioFormat;
+    input_audio_transcription: InputAudioTranscription | null;
+    turn_detection: TurnDetectionType | null;
+    tools: Tool[];
+    tool_choice: ToolChoice;
+    temperature: number;
+    max_response_output_tokens: number | 'inf';
+    expires_at: number;
+}
+export interface ConversationResource {
+    id: string;
+    object: 'realtime.conversation';
+}
+export type ResponseStatusDetails = {
+    type: 'incomplete';
+    reason: 'max_output_tokens' | 'content_filter' | string;
+} | {
+    type: 'failed';
+    error?: {
+        code: 'server_error' | 'rate_limit_exceeded' | string;
+        message: string;
+    };
+} | {
+    type: 'cancelled';
+    reason: 'turn_detected' | 'client_cancelled' | string;
+};
+export interface ModelUsage {
+    total_tokens: number;
+    input_tokens: number;
+    output_tokens: number;
+    input_token_details: {
+        text_tokens: number;
+        audio_tokens: number;
+        cached_tokens: number;
+        cached_tokens_details: {
+            text_tokens: number;
+            audio_tokens: number;
+        };
+    };
+    output_token_details: {
+        text_tokens: number;
+        audio_tokens: number;
+    };
+}
+export interface ResponseResource {
+    id: string;
+    object: 'realtime.response';
+    status: ResponseStatus;
+    status_details: ResponseStatusDetails;
+    output: ItemResource[];
+    usage?: ModelUsage;
+}
+interface BaseClientEvent {
+    event_id?: string;
+    type: ClientEventType;
+}
+export interface SessionUpdateEvent extends BaseClientEvent {
+    type: 'session.update';
+    session: Partial<{
+        modalities: ['text', 'audio'] | ['text'];
+        instructions: string;
+        voice: Voice;
+        input_audio_format: AudioFormat;
+        output_audio_format: AudioFormat;
+        input_audio_transcription: InputAudioTranscription | null;
+        turn_detection: TurnDetectionType | null;
+        tools: Tool[];
+        tool_choice: ToolChoice;
+        temperature: number;
+        max_response_output_tokens?: number | 'inf';
+    }>;
+}
+export interface InputAudioBufferAppendEvent extends BaseClientEvent {
+    type: 'input_audio_buffer.append';
+    audio: AudioBase64Bytes;
+}
+export interface InputAudioBufferCommitEvent extends BaseClientEvent {
+    type: 'input_audio_buffer.commit';
+}
+export interface InputAudioBufferClearEvent extends BaseClientEvent {
+    type: 'input_audio_buffer.clear';
+}
+export interface UserItemCreate {
+    type: 'message';
+    role: 'user';
+    content: (InputTextContent | InputAudioContent)[];
+}
+export interface AssistantItemCreate {
+    type: 'message';
+    role: 'assistant';
+    content: TextContent[];
+}
+export interface SystemItemCreate {
+    type: 'message';
+    role: 'system';
+    content: InputTextContent[];
+}
+export interface FunctionCallOutputItemCreate {
+    type: 'function_call_output';
+    call_id: string;
+    output: string;
+}
+export type ConversationItemCreateContent = UserItemCreate | AssistantItemCreate | SystemItemCreate | FunctionCallOutputItemCreate;
+export interface ConversationItemCreateEvent extends BaseClientEvent {
+    type: 'conversation.item.create';
+    previous_item_id?: string;
+    item: ConversationItemCreateContent;
+}
+export interface ConversationItemTruncateEvent extends BaseClientEvent {
+    type: 'conversation.item.truncate';
+    item_id: string;
+    content_index: number;
+    audio_end_ms: number;
+}
+export interface ConversationItemDeleteEvent extends BaseClientEvent {
+    type: 'conversation.item.delete';
+    item_id: string;
+}
+export interface ResponseCreateEvent extends BaseClientEvent {
+    type: 'response.create';
+    response?: Partial<{
+        modalities: ['text', 'audio'] | ['text'];
+        instructions: string;
+        voice: Voice;
+        output_audio_format: AudioFormat;
+        tools?: Tool[];
+        tool_choice: ToolChoice;
+        temperature: number;
+        max_output_tokens: number | 'inf';
+    }>;
+}
+export interface ResponseCancelEvent extends BaseClientEvent {
+    type: 'response.cancel';
+}
+export type ClientEvent = SessionUpdateEvent | InputAudioBufferAppendEvent | InputAudioBufferCommitEvent | InputAudioBufferClearEvent | ConversationItemCreateEvent | ConversationItemTruncateEvent | ConversationItemDeleteEvent | ResponseCreateEvent | ResponseCancelEvent;
+interface BaseServerEvent {
+    event_id: string;
+    type: ServerEventType;
+}
+export interface ErrorEvent extends BaseServerEvent {
+    type: 'error';
+    error: {
+        type: 'invalid_request_error' | 'server_error' | string;
+        code?: string;
+        message: string;
+        param: string;
+        event_id: string;
+    };
+}
+export interface SessionCreatedEvent extends BaseServerEvent {
+    type: 'session.created';
+    session: SessionResource;
+}
+export interface SessionUpdatedEvent extends BaseServerEvent {
+    type: 'session.updated';
+    session: SessionResource;
+}
+export interface ConversationCreatedEvent extends BaseServerEvent {
+    type: 'conversation.created';
+    conversation: ConversationResource;
+}
+export interface InputAudioBufferCommittedEvent extends BaseServerEvent {
+    type: 'input_audio_buffer.committed';
+    item_id: string;
+}
+export interface InputAudioBufferClearedEvent extends BaseServerEvent {
+    type: 'input_audio_buffer.cleared';
+}
+export interface InputAudioBufferSpeechStartedEvent extends BaseServerEvent {
+    type: 'input_audio_buffer.speech_started';
+    audio_start_ms: number;
+    item_id: string;
+}
+export interface InputAudioBufferSpeechStoppedEvent extends BaseServerEvent {
+    type: 'input_audio_buffer.speech_stopped';
+    audio_end_ms: number;
+    item_id: string;
+}
+export interface ConversationItemCreatedEvent extends BaseServerEvent {
+    type: 'conversation.item.created';
+    item: ItemResource;
+}
+export interface ConversationItemInputAudioTranscriptionCompletedEvent extends BaseServerEvent {
+    type: 'conversation.item.input_audio_transcription.completed';
+    item_id: string;
+    content_index: number;
+    transcript: string;
+}
+export interface ConversationItemInputAudioTranscriptionFailedEvent extends BaseServerEvent {
+    type: 'conversation.item.input_audio_transcription.failed';
+    item_id: string;
+    content_index: number;
+    error: {
+        type: string;
+        code?: string;
+        message: string;
+        param: null;
+    };
+}
+export interface ConversationItemTruncatedEvent extends BaseServerEvent {
+    type: 'conversation.item.truncated';
+    item_id: string;
+    content_index: number;
+    audio_end_ms: number;
+}
+export interface ConversationItemDeletedEvent extends BaseServerEvent {
+    type: 'conversation.item.deleted';
+    item_id: string;
+}
+export interface ResponseCreatedEvent extends BaseServerEvent {
+    type: 'response.created';
+    response: ResponseResource;
+}
+export interface ResponseDoneEvent extends BaseServerEvent {
+    type: 'response.done';
+    response: ResponseResource;
+}
+export interface ResponseOutputItemAddedEvent extends BaseServerEvent {
+    type: 'response.output_item.added';
+    response_id: string;
+    output_index: number;
+    item: ItemResource;
+}
+export interface ResponseOutputItemDoneEvent extends BaseServerEvent {
+    type: 'response.output_item.done';
+    response_id: string;
+    output_index: number;
+    item: ItemResource;
+}
+export interface ResponseContentPartAddedEvent extends BaseServerEvent {
+    type: 'response.content_part.added';
+    response_id: string;
+    item_id: string;
+    output_index: number;
+    content_index: number;
+    part: ContentPart;
+}
+export interface ResponseContentPartDoneEvent extends BaseServerEvent {
+    type: 'response.content_part.done';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    part: ContentPart;
+}
+export interface ResponseTextDeltaEvent extends BaseServerEvent {
+    type: 'response.text.delta';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    delta: string;
+}
+export interface ResponseTextDoneEvent extends BaseServerEvent {
+    type: 'response.text.done';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    text: string;
+}
+export interface ResponseAudioTranscriptDeltaEvent extends BaseServerEvent {
+    type: 'response.audio_transcript.delta';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    delta: string;
+}
+export interface ResponseAudioTranscriptDoneEvent extends BaseServerEvent {
+    type: 'response.audio_transcript.done';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    transcript: string;
+}
+export interface ResponseAudioDeltaEvent extends BaseServerEvent {
+    type: 'response.audio.delta';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+    delta: AudioBase64Bytes;
+}
+export interface ResponseAudioDoneEvent extends BaseServerEvent {
+    type: 'response.audio.done';
+    response_id: string;
+    output_index: number;
+    content_index: number;
+}
+export interface ResponseFunctionCallArgumentsDeltaEvent extends BaseServerEvent {
+    type: 'response.function_call_arguments.delta';
+    response_id: string;
+    output_index: number;
+    delta: string;
+}
+export interface ResponseFunctionCallArgumentsDoneEvent extends BaseServerEvent {
+    type: 'response.function_call_arguments.done';
+    response_id: string;
+    output_index: number;
+    arguments: string;
+}
+export interface RateLimitsUpdatedEvent extends BaseServerEvent {
+    type: 'rate_limits.updated';
+    rate_limits: {
+        name: 'requests' | 'tokens' | 'input_tokens' | 'output_tokens' | string;
+        limit: number;
+        remaining: number;
+        reset_seconds: number;
+    }[];
+}
+export type ServerEvent = ErrorEvent | SessionCreatedEvent | SessionUpdatedEvent | ConversationCreatedEvent | InputAudioBufferCommittedEvent | InputAudioBufferClearedEvent | InputAudioBufferSpeechStartedEvent | InputAudioBufferSpeechStoppedEvent | ConversationItemCreatedEvent | ConversationItemInputAudioTranscriptionCompletedEvent | ConversationItemInputAudioTranscriptionFailedEvent | ConversationItemTruncatedEvent | ConversationItemDeletedEvent | ResponseCreatedEvent | ResponseDoneEvent | ResponseOutputItemAddedEvent | ResponseOutputItemDoneEvent | ResponseContentPartAddedEvent | ResponseContentPartDoneEvent | ResponseTextDeltaEvent | ResponseTextDoneEvent | ResponseAudioTranscriptDeltaEvent | ResponseAudioTranscriptDoneEvent | ResponseAudioDeltaEvent | ResponseAudioDoneEvent | ResponseFunctionCallArgumentsDeltaEvent | ResponseFunctionCallArgumentsDoneEvent | RateLimitsUpdatedEvent;
+export {};
+//# sourceMappingURL=api_proto.d.ts.map

package/dist/realtime/index.d.cts

@@ -0,0 +1,3 @@
+export * from './api_proto.js';
+export * from './realtime_model.js';
+//# sourceMappingURL=index.d.ts.map

package/dist/realtime/realtime_model.cjs

@@ -276,7 +276,7 @@ class RealtimeModel extends import_agents.multimodal.RealtimeModel {
     entraToken = void 0
   }) {
     super();
-    if (apiKey === "") {
+    if (apiKey === "" && !(isAzure && entraToken)) {
       throw new Error(
         "OpenAI API key is required, either using the argument or by setting the OPENAI_API_KEY environmental variable"
       );