@tanstack/ai-grok 0.6.8 → 0.7.1

package/src/index.ts

@@ -33,6 +33,31 @@ export type {
   GrokImageModelProviderOptionsByName,
 } from './image/image-provider-options'
 
+// Speech (TTS) adapter - for text-to-speech
+export {
+  GrokSpeechAdapter,
+  createGrokSpeech,
+  grokSpeech,
+  type GrokSpeechConfig,
+} from './adapters/tts'
+export type {
+  GrokTTSProviderOptions,
+  GrokTTSVoice,
+  GrokTTSCodec,
+} from './audio/tts-provider-options'
+
+// Transcription adapter - for speech-to-text
+export {
+  GrokTranscriptionAdapter,
+  createGrokTranscription,
+  grokTranscription,
+  type GrokTranscriptionConfig,
+} from './adapters/transcription'
+export type {
+  GrokTranscriptionProviderOptions,
+  GrokSTTAudioFormat,
+} from './audio/transcription-provider-options'
+
 // ============================================================================
 // Type Exports
 // ============================================================================
@@ -45,8 +70,17 @@ export type {
   ResolveInputModalities,
   GrokChatModel,
   GrokImageModel,
+  GrokTTSModel,
+  GrokTranscriptionModel,
+  GrokRealtimeModel,
+} from './model-meta'
+export {
+  GROK_CHAT_MODELS,
+  GROK_IMAGE_MODELS,
+  GROK_TTS_MODELS,
+  GROK_TRANSCRIPTION_MODELS,
+  GROK_REALTIME_MODELS,
 } from './model-meta'
-export { GROK_CHAT_MODELS, GROK_IMAGE_MODELS } from './model-meta'
 export type {
   GrokTextMetadata,
   GrokImageMetadata,
@@ -55,3 +89,18 @@ export type {
   GrokDocumentMetadata,
   GrokMessageMetadataByModality,
 } from './message-types'
+
+// ============================================================================
+// Realtime (Voice Agent) Adapters
+// ============================================================================
+
+export { grokRealtimeToken, grokRealtime } from './realtime/index'
+
+export type {
+  GrokRealtimeVoice,
+  GrokRealtimeTokenOptions,
+  GrokRealtimeOptions,
+  GrokTurnDetection,
+  GrokSemanticVADConfig,
+  GrokServerVADConfig,
+} from './realtime/index'

package/src/model-meta.ts

@@ -283,8 +283,62 @@ export const GROK_CHAT_MODELS = [
  */
 export const GROK_IMAGE_MODELS = [GROK_2_IMAGE.name] as const
 
+// xAI's `/v1/tts` endpoint is endpoint-addressed and does not take a `model`
+// parameter. This synthetic identifier satisfies the SDK's `TTSOptions.model`
+// contract and provides a stable value for logging and fixture matching.
+const GROK_TTS = {
+  name: 'grok-tts',
+  supports: {
+    input: ['text'],
+    output: ['audio'],
+  },
+} as const satisfies ModelMeta
+
+// xAI's `/v1/stt` endpoint is endpoint-addressed and does not take a `model`
+// parameter. This synthetic identifier satisfies the SDK's
+// `TranscriptionOptions.model` contract.
+const GROK_STT = {
+  name: 'grok-stt',
+  supports: {
+    input: ['audio'],
+    output: ['text'],
+  },
+} as const satisfies ModelMeta
+
+const GROK_VOICE_FAST_1 = {
+  name: 'grok-voice-fast-1.0',
+  supports: {
+    input: ['audio', 'text'],
+    output: ['audio', 'text'],
+    capabilities: ['tool_calling'],
+    tools: [] as const,
+  },
+} as const satisfies ModelMeta
+
+const GROK_VOICE_THINK_FAST_1 = {
+  name: 'grok-voice-think-fast-1.0',
+  supports: {
+    input: ['audio', 'text'],
+    output: ['audio', 'text'],
+    capabilities: ['reasoning', 'tool_calling'],
+    tools: [] as const,
+  },
+} as const satisfies ModelMeta
+
+export const GROK_TTS_MODELS = [GROK_TTS.name] as const
+
+export const GROK_TRANSCRIPTION_MODELS = [GROK_STT.name] as const
+
+export const GROK_REALTIME_MODELS = [
+  GROK_VOICE_FAST_1.name,
+  GROK_VOICE_THINK_FAST_1.name,
+] as const
+
 export type GrokChatModel = (typeof GROK_CHAT_MODELS)[number]
 export type GrokImageModel = (typeof GROK_IMAGE_MODELS)[number]
+export type GrokTTSModel = (typeof GROK_TTS_MODELS)[number]
+export type GrokTranscriptionModel = (typeof GROK_TRANSCRIPTION_MODELS)[number]
+export type GrokRealtimeModel = (typeof GROK_REALTIME_MODELS)[number]
 
 /**
  * Type-only map from Grok chat model name to its supported input modalities.