@ai-sdk/openai 1.3.10 → 1.3.12

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (14) hide show
  1. package/CHANGELOG.md +15 -0
  2. package/dist/index.d.mts +8 -2
  3. package/dist/index.d.ts +8 -2
  4. package/dist/index.js +112 -5
  5. package/dist/index.js.map +1 -1
  6. package/dist/index.mjs +114 -2
  7. package/dist/index.mjs.map +1 -1
  8. package/internal/dist/index.d.mts +35 -7
  9. package/internal/dist/index.d.ts +35 -7
  10. package/internal/dist/index.js +213 -112
  11. package/internal/dist/index.js.map +1 -1
  12. package/internal/dist/index.mjs +213 -108
  13. package/internal/dist/index.mjs.map +1 -1
  14. package/package.json +3 -3
package/internal/dist/index.d.mts CHANGED
@@ -1,8 +1,8 @@
1
- import { LanguageModelV1, EmbeddingModelV1, ImageModelV1, TranscriptionModelV1CallOptions, TranscriptionModelV1 } from '@ai-sdk/provider';
1
+ import { LanguageModelV1, EmbeddingModelV1, ImageModelV1, TranscriptionModelV1CallOptions, TranscriptionModelV1, SpeechModelV1 } from '@ai-sdk/provider';
2
2
  import { FetchFunction } from '@ai-sdk/provider-utils';
3
3
  import { z } from 'zod';
4
4
 
5
- type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
5
+ type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4.1' | 'gpt-4.1-2025-04-14' | 'gpt-4.1-mini' | 'gpt-4.1-mini-2025-04-14' | 'gpt-4.1-nano' | 'gpt-4.1-nano-2025-04-14' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
6
6
  interface OpenAIChatSettings {
7
7
  /**
8
8
  Modify the likelihood of specified tokens appearing in the completion.
@@ -270,7 +270,7 @@ type OpenAITranscriptionModelOptions = {
270
270
  timestamp_granularities?: Array<'word' | 'segment'>;
271
271
  };
272
272
 
273
- declare const OpenAIProviderOptionsSchema: z.ZodObject<{
273
+ declare const openAIProviderOptionsSchema: z.ZodObject<{
274
274
  include: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodString, "many">>>;
275
275
  language: z.ZodOptional<z.ZodNullable<z.ZodString>>;
276
276
  prompt: z.ZodOptional<z.ZodNullable<z.ZodString>>;
@@ -291,7 +291,7 @@ declare const OpenAIProviderOptionsSchema: z.ZodObject<{
291
291
  }>;
292
292
  type OpenAITranscriptionCallOptions = Omit<TranscriptionModelV1CallOptions, 'providerOptions'> & {
293
293
  providerOptions?: {
294
- openai?: z.infer<typeof OpenAIProviderOptionsSchema>;
294
+ openai?: z.infer<typeof openAIProviderOptionsSchema>;
295
295
  };
296
296
  };
297
297
  interface OpenAITranscriptionModelConfig extends OpenAIConfig {
@@ -309,6 +309,34 @@ declare class OpenAITranscriptionModel implements TranscriptionModelV1 {
309
309
  doGenerate(options: OpenAITranscriptionCallOptions): Promise<Awaited<ReturnType<TranscriptionModelV1['doGenerate']>>>;
310
310
  }
311
311
 
312
+ type OpenAISpeechModelId = 'tts-1' | 'tts-1-hd' | 'gpt-4o-mini-tts' | (string & {});
313
+
314
+ declare const OpenAIProviderOptionsSchema: z.ZodObject<{
315
+ instructions: z.ZodOptional<z.ZodNullable<z.ZodString>>;
316
+ speed: z.ZodOptional<z.ZodNullable<z.ZodDefault<z.ZodNumber>>>;
317
+ }, "strip", z.ZodTypeAny, {
318
+ instructions?: string | null | undefined;
319
+ speed?: number | null | undefined;
320
+ }, {
321
+ instructions?: string | null | undefined;
322
+ speed?: number | null | undefined;
323
+ }>;
324
+ type OpenAISpeechCallOptions = z.infer<typeof OpenAIProviderOptionsSchema>;
325
+ interface OpenAISpeechModelConfig extends OpenAIConfig {
326
+ _internal?: {
327
+ currentDate?: () => Date;
328
+ };
329
+ }
330
+ declare class OpenAISpeechModel implements SpeechModelV1 {
331
+ readonly modelId: OpenAISpeechModelId;
332
+ private readonly config;
333
+ readonly specificationVersion = "v1";
334
+ get provider(): string;
335
+ constructor(modelId: OpenAISpeechModelId, config: OpenAISpeechModelConfig);
336
+ private getArgs;
337
+ doGenerate(options: Parameters<SpeechModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<SpeechModelV1['doGenerate']>>>;
338
+ }
339
+
312
340
  type OpenAIResponsesModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
313
341
 
314
342
  declare class OpenAIResponsesLanguageModel implements LanguageModelV1 {
@@ -336,20 +364,20 @@ declare const openaiResponsesProviderOptionsSchema: z.ZodObject<{
336
364
  store?: boolean | null | undefined;
337
365
  metadata?: any;
338
366
  reasoningEffort?: string | null | undefined;
367
+ instructions?: string | null | undefined;
339
368
  parallelToolCalls?: boolean | null | undefined;
340
369
  previousResponseId?: string | null | undefined;
341
370
  strictSchemas?: boolean | null | undefined;
342
- instructions?: string | null | undefined;
343
371
  }, {
344
372
  user?: string | null | undefined;
345
373
  store?: boolean | null | undefined;
346
374
  metadata?: any;
347
375
  reasoningEffort?: string | null | undefined;
376
+ instructions?: string | null | undefined;
348
377
  parallelToolCalls?: boolean | null | undefined;
349
378
  previousResponseId?: string | null | undefined;
350
379
  strictSchemas?: boolean | null | undefined;
351
- instructions?: string | null | undefined;
352
380
  }>;
353
381
  type OpenAIResponsesProviderOptions = z.infer<typeof openaiResponsesProviderOptionsSchema>;
354
382
 
355
- export { OpenAIChatLanguageModel, type OpenAIChatModelId, type OpenAIChatSettings, OpenAICompletionLanguageModel, type OpenAICompletionModelId, type OpenAICompletionSettings, OpenAIEmbeddingModel, type OpenAIEmbeddingModelId, type OpenAIEmbeddingSettings, OpenAIImageModel, type OpenAIImageModelId, type OpenAIImageSettings, OpenAIResponsesLanguageModel, type OpenAIResponsesProviderOptions, type OpenAITranscriptionCallOptions, OpenAITranscriptionModel, type OpenAITranscriptionModelId, type OpenAITranscriptionModelOptions, modelMaxImagesPerCall };
383
+ export { OpenAIChatLanguageModel, type OpenAIChatModelId, type OpenAIChatSettings, OpenAICompletionLanguageModel, type OpenAICompletionModelId, type OpenAICompletionSettings, OpenAIEmbeddingModel, type OpenAIEmbeddingModelId, type OpenAIEmbeddingSettings, OpenAIImageModel, type OpenAIImageModelId, type OpenAIImageSettings, OpenAIResponsesLanguageModel, type OpenAIResponsesProviderOptions, type OpenAISpeechCallOptions, OpenAISpeechModel, type OpenAISpeechModelId, type OpenAITranscriptionCallOptions, OpenAITranscriptionModel, type OpenAITranscriptionModelId, type OpenAITranscriptionModelOptions, modelMaxImagesPerCall };
package/internal/dist/index.d.ts CHANGED
@@ -1,8 +1,8 @@
1
- import { LanguageModelV1, EmbeddingModelV1, ImageModelV1, TranscriptionModelV1CallOptions, TranscriptionModelV1 } from '@ai-sdk/provider';
1
+ import { LanguageModelV1, EmbeddingModelV1, ImageModelV1, TranscriptionModelV1CallOptions, TranscriptionModelV1, SpeechModelV1 } from '@ai-sdk/provider';
2
2
  import { FetchFunction } from '@ai-sdk/provider-utils';
3
3
  import { z } from 'zod';
4
4
 
5
- type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
5
+ type OpenAIChatModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4.1' | 'gpt-4.1-2025-04-14' | 'gpt-4.1-mini' | 'gpt-4.1-mini-2025-04-14' | 'gpt-4.1-nano' | 'gpt-4.1-nano-2025-04-14' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | 'gpt-4o-audio-preview' | 'gpt-4o-audio-preview-2024-10-01' | 'gpt-4o-audio-preview-2024-12-17' | 'gpt-4o-search-preview' | 'gpt-4o-search-preview-2025-03-11' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
6
6
  interface OpenAIChatSettings {
7
7
  /**
8
8
  Modify the likelihood of specified tokens appearing in the completion.
@@ -270,7 +270,7 @@ type OpenAITranscriptionModelOptions = {
270
270
  timestamp_granularities?: Array<'word' | 'segment'>;
271
271
  };
272
272
 
273
- declare const OpenAIProviderOptionsSchema: z.ZodObject<{
273
+ declare const openAIProviderOptionsSchema: z.ZodObject<{
274
274
  include: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodString, "many">>>;
275
275
  language: z.ZodOptional<z.ZodNullable<z.ZodString>>;
276
276
  prompt: z.ZodOptional<z.ZodNullable<z.ZodString>>;
@@ -291,7 +291,7 @@ declare const OpenAIProviderOptionsSchema: z.ZodObject<{
291
291
  }>;
292
292
  type OpenAITranscriptionCallOptions = Omit<TranscriptionModelV1CallOptions, 'providerOptions'> & {
293
293
  providerOptions?: {
294
- openai?: z.infer<typeof OpenAIProviderOptionsSchema>;
294
+ openai?: z.infer<typeof openAIProviderOptionsSchema>;
295
295
  };
296
296
  };
297
297
  interface OpenAITranscriptionModelConfig extends OpenAIConfig {
@@ -309,6 +309,34 @@ declare class OpenAITranscriptionModel implements TranscriptionModelV1 {
309
309
  doGenerate(options: OpenAITranscriptionCallOptions): Promise<Awaited<ReturnType<TranscriptionModelV1['doGenerate']>>>;
310
310
  }
311
311
 
312
+ type OpenAISpeechModelId = 'tts-1' | 'tts-1-hd' | 'gpt-4o-mini-tts' | (string & {});
313
+
314
+ declare const OpenAIProviderOptionsSchema: z.ZodObject<{
315
+ instructions: z.ZodOptional<z.ZodNullable<z.ZodString>>;
316
+ speed: z.ZodOptional<z.ZodNullable<z.ZodDefault<z.ZodNumber>>>;
317
+ }, "strip", z.ZodTypeAny, {
318
+ instructions?: string | null | undefined;
319
+ speed?: number | null | undefined;
320
+ }, {
321
+ instructions?: string | null | undefined;
322
+ speed?: number | null | undefined;
323
+ }>;
324
+ type OpenAISpeechCallOptions = z.infer<typeof OpenAIProviderOptionsSchema>;
325
+ interface OpenAISpeechModelConfig extends OpenAIConfig {
326
+ _internal?: {
327
+ currentDate?: () => Date;
328
+ };
329
+ }
330
+ declare class OpenAISpeechModel implements SpeechModelV1 {
331
+ readonly modelId: OpenAISpeechModelId;
332
+ private readonly config;
333
+ readonly specificationVersion = "v1";
334
+ get provider(): string;
335
+ constructor(modelId: OpenAISpeechModelId, config: OpenAISpeechModelConfig);
336
+ private getArgs;
337
+ doGenerate(options: Parameters<SpeechModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<SpeechModelV1['doGenerate']>>>;
338
+ }
339
+
312
340
  type OpenAIResponsesModelId = 'o1' | 'o1-2024-12-17' | 'o1-mini' | 'o1-mini-2024-09-12' | 'o1-preview' | 'o1-preview-2024-09-12' | 'o3-mini' | 'o3-mini-2025-01-31' | 'gpt-4o' | 'gpt-4o-2024-05-13' | 'gpt-4o-2024-08-06' | 'gpt-4o-2024-11-20' | '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' | 'gpt-4-0613' | 'gpt-4.5-preview' | 'gpt-4.5-preview-2025-02-27' | 'gpt-3.5-turbo-0125' | 'gpt-3.5-turbo' | 'gpt-3.5-turbo-1106' | 'chatgpt-4o-latest' | (string & {});
313
341
 
314
342
  declare class OpenAIResponsesLanguageModel implements LanguageModelV1 {
@@ -336,20 +364,20 @@ declare const openaiResponsesProviderOptionsSchema: z.ZodObject<{
336
364
  store?: boolean | null | undefined;
337
365
  metadata?: any;
338
366
  reasoningEffort?: string | null | undefined;
367
+ instructions?: string | null | undefined;
339
368
  parallelToolCalls?: boolean | null | undefined;
340
369
  previousResponseId?: string | null | undefined;
341
370
  strictSchemas?: boolean | null | undefined;
342
- instructions?: string | null | undefined;
343
371
  }, {
344
372
  user?: string | null | undefined;
345
373
  store?: boolean | null | undefined;
346
374
  metadata?: any;
347
375
  reasoningEffort?: string | null | undefined;
376
+ instructions?: string | null | undefined;
348
377
  parallelToolCalls?: boolean | null | undefined;
349
378
  previousResponseId?: string | null | undefined;
350
379
  strictSchemas?: boolean | null | undefined;
351
- instructions?: string | null | undefined;
352
380
  }>;
353
381
  type OpenAIResponsesProviderOptions = z.infer<typeof openaiResponsesProviderOptionsSchema>;
354
382
 
355
- export { OpenAIChatLanguageModel, type OpenAIChatModelId, type OpenAIChatSettings, OpenAICompletionLanguageModel, type OpenAICompletionModelId, type OpenAICompletionSettings, OpenAIEmbeddingModel, type OpenAIEmbeddingModelId, type OpenAIEmbeddingSettings, OpenAIImageModel, type OpenAIImageModelId, type OpenAIImageSettings, OpenAIResponsesLanguageModel, type OpenAIResponsesProviderOptions, type OpenAITranscriptionCallOptions, OpenAITranscriptionModel, type OpenAITranscriptionModelId, type OpenAITranscriptionModelOptions, modelMaxImagesPerCall };
383
+ export { OpenAIChatLanguageModel, type OpenAIChatModelId, type OpenAIChatSettings, OpenAICompletionLanguageModel, type OpenAICompletionModelId, type OpenAICompletionSettings, OpenAIEmbeddingModel, type OpenAIEmbeddingModelId, type OpenAIEmbeddingSettings, OpenAIImageModel, type OpenAIImageModelId, type OpenAIImageSettings, OpenAIResponsesLanguageModel, type OpenAIResponsesProviderOptions, type OpenAISpeechCallOptions, OpenAISpeechModel, type OpenAISpeechModelId, type OpenAITranscriptionCallOptions, OpenAITranscriptionModel, type OpenAITranscriptionModelId, type OpenAITranscriptionModelOptions, modelMaxImagesPerCall };
package/internal/dist/index.js CHANGED
@@ -25,6 +25,7 @@ __export(internal_exports, {
25
25
  OpenAIEmbeddingModel: () => OpenAIEmbeddingModel,
26
26
  OpenAIImageModel: () => OpenAIImageModel,
27
27
  OpenAIResponsesLanguageModel: () => OpenAIResponsesLanguageModel,
28
+ OpenAISpeechModel: () => OpenAISpeechModel,
28
29
  OpenAITranscriptionModel: () => OpenAITranscriptionModel,
29
30
  modelMaxImagesPerCall: () => modelMaxImagesPerCall
30
31
  });
@@ -1615,7 +1616,7 @@ var openaiImageResponseSchema = import_zod5.z.object({
1615
1616
  // src/openai-transcription-model.ts
1616
1617
  var import_provider_utils7 = require("@ai-sdk/provider-utils");
1617
1618
  var import_zod6 = require("zod");
1618
- var OpenAIProviderOptionsSchema = import_zod6.z.object({
1619
+ var openAIProviderOptionsSchema = import_zod6.z.object({
1619
1620
  include: import_zod6.z.array(import_zod6.z.string()).nullish(),
1620
1621
  language: import_zod6.z.string().nullish(),
1621
1622
  prompt: import_zod6.z.string().nullish(),
@@ -1700,7 +1701,7 @@ var OpenAITranscriptionModel = class {
1700
1701
  const openAIOptions = (0, import_provider_utils7.parseProviderOptions)({
1701
1702
  provider: "openai",
1702
1703
  providerOptions,
1703
- schema: OpenAIProviderOptionsSchema
1704
+ schema: openAIProviderOptionsSchema
1704
1705
  });
1705
1706
  const formData = new FormData();
1706
1707
  const blob = audio instanceof Uint8Array ? new Blob([audio]) : new Blob([(0, import_provider_utils7.convertBase64ToUint8Array)(audio)]);
@@ -1781,13 +1782,112 @@ var openaiTranscriptionResponseSchema = import_zod6.z.object({
1781
1782
  ).nullish()
1782
1783
  });
1783
1784
 
1784
- // src/responses/openai-responses-language-model.ts
1785
- var import_provider_utils9 = require("@ai-sdk/provider-utils");
1785
+ // src/openai-speech-model.ts
1786
+ var import_provider_utils8 = require("@ai-sdk/provider-utils");
1786
1787
  var import_zod7 = require("zod");
1788
+ var OpenAIProviderOptionsSchema = import_zod7.z.object({
1789
+ instructions: import_zod7.z.string().nullish(),
1790
+ speed: import_zod7.z.number().min(0.25).max(4).default(1).nullish()
1791
+ });
1792
+ var OpenAISpeechModel = class {
1793
+ constructor(modelId, config) {
1794
+ this.modelId = modelId;
1795
+ this.config = config;
1796
+ this.specificationVersion = "v1";
1797
+ }
1798
+ get provider() {
1799
+ return this.config.provider;
1800
+ }
1801
+ getArgs({
1802
+ text,
1803
+ voice = "alloy",
1804
+ outputFormat = "mp3",
1805
+ speed,
1806
+ instructions,
1807
+ providerOptions
1808
+ }) {
1809
+ const warnings = [];
1810
+ const openAIOptions = (0, import_provider_utils8.parseProviderOptions)({
1811
+ provider: "openai",
1812
+ providerOptions,
1813
+ schema: OpenAIProviderOptionsSchema
1814
+ });
1815
+ const requestBody = {
1816
+ model: this.modelId,
1817
+ input: text,
1818
+ voice,
1819
+ response_format: "mp3",
1820
+ speed,
1821
+ instructions
1822
+ };
1823
+ if (outputFormat) {
1824
+ if (["mp3", "opus", "aac", "flac", "wav", "pcm"].includes(outputFormat)) {
1825
+ requestBody.response_format = outputFormat;
1826
+ } else {
1827
+ warnings.push({
1828
+ type: "unsupported-setting",
1829
+ setting: "outputFormat",
1830
+ details: `Unsupported output format: ${outputFormat}. Using mp3 instead.`
1831
+ });
1832
+ }
1833
+ }
1834
+ if (openAIOptions) {
1835
+ const speechModelOptions = {};
1836
+ for (const key in speechModelOptions) {
1837
+ const value = speechModelOptions[key];
1838
+ if (value !== void 0) {
1839
+ requestBody[key] = value;
1840
+ }
1841
+ }
1842
+ }
1843
+ return {
1844
+ requestBody,
1845
+ warnings
1846
+ };
1847
+ }
1848
+ async doGenerate(options) {
1849
+ var _a, _b, _c;
1850
+ const currentDate = (_c = (_b = (_a = this.config._internal) == null ? void 0 : _a.currentDate) == null ? void 0 : _b.call(_a)) != null ? _c : /* @__PURE__ */ new Date();
1851
+ const { requestBody, warnings } = this.getArgs(options);
1852
+ const {
1853
+ value: audio,
1854
+ responseHeaders,
1855
+ rawValue: rawResponse
1856
+ } = await (0, import_provider_utils8.postJsonToApi)({
1857
+ url: this.config.url({
1858
+ path: "/audio/speech",
1859
+ modelId: this.modelId
1860
+ }),
1861
+ headers: (0, import_provider_utils8.combineHeaders)(this.config.headers(), options.headers),
1862
+ body: requestBody,
1863
+ failedResponseHandler: openaiFailedResponseHandler,
1864
+ successfulResponseHandler: (0, import_provider_utils8.createBinaryResponseHandler)(),
1865
+ abortSignal: options.abortSignal,
1866
+ fetch: this.config.fetch
1867
+ });
1868
+ return {
1869
+ audio,
1870
+ warnings,
1871
+ request: {
1872
+ body: JSON.stringify(requestBody)
1873
+ },
1874
+ response: {
1875
+ timestamp: currentDate,
1876
+ modelId: this.modelId,
1877
+ headers: responseHeaders,
1878
+ body: rawResponse
1879
+ }
1880
+ };
1881
+ }
1882
+ };
1883
+
1884
+ // src/responses/openai-responses-language-model.ts
1885
+ var import_provider_utils10 = require("@ai-sdk/provider-utils");
1886
+ var import_zod8 = require("zod");
1787
1887
 
1788
1888
  // src/responses/convert-to-openai-responses-messages.ts
1789
1889
  var import_provider7 = require("@ai-sdk/provider");
1790
- var import_provider_utils8 = require("@ai-sdk/provider-utils");
1890
+ var import_provider_utils9 = require("@ai-sdk/provider-utils");
1791
1891
  function convertToOpenAIResponsesMessages({
1792
1892
  prompt,
1793
1893
  systemMessageMode
@@ -1834,7 +1934,7 @@ function convertToOpenAIResponsesMessages({
1834
1934
  case "image": {
1835
1935
  return {
1836
1936
  type: "input_image",
1837
- image_url: part.image instanceof URL ? part.image.toString() : `data:${(_a = part.mimeType) != null ? _a : "image/jpeg"};base64,${(0, import_provider_utils8.convertUint8ArrayToBase64)(part.image)}`,
1937
+ image_url: part.image instanceof URL ? part.image.toString() : `data:${(_a = part.mimeType) != null ? _a : "image/jpeg"};base64,${(0, import_provider_utils9.convertUint8ArrayToBase64)(part.image)}`,
1838
1938
  // OpenAI specific extension: image detail
1839
1939
  detail: (_c = (_b = part.providerMetadata) == null ? void 0 : _b.openai) == null ? void 0 : _c.imageDetail
1840
1940
  };
@@ -2070,7 +2170,7 @@ var OpenAIResponsesLanguageModel = class {
2070
2170
  systemMessageMode: modelConfig.systemMessageMode
2071
2171
  });
2072
2172
  warnings.push(...messageWarnings);
2073
- const openaiOptions = (0, import_provider_utils9.parseProviderOptions)({
2173
+ const openaiOptions = (0, import_provider_utils10.parseProviderOptions)({
2074
2174
  provider: "openai",
2075
2175
  providerOptions: providerMetadata,
2076
2176
  schema: openaiResponsesProviderOptionsSchema
@@ -2190,58 +2290,58 @@ var OpenAIResponsesLanguageModel = class {
2190
2290
  responseHeaders,
2191
2291
  value: response,
2192
2292
  rawValue: rawResponse
2193
- } = await (0, import_provider_utils9.postJsonToApi)({
2293
+ } = await (0, import_provider_utils10.postJsonToApi)({
2194
2294
  url: this.config.url({
2195
2295
  path: "/responses",
2196
2296
  modelId: this.modelId
2197
2297
  }),
2198
- headers: (0, import_provider_utils9.combineHeaders)(this.config.headers(), options.headers),
2298
+ headers: (0, import_provider_utils10.combineHeaders)(this.config.headers(), options.headers),
2199
2299
  body,
2200
2300
  failedResponseHandler: openaiFailedResponseHandler,
2201
- successfulResponseHandler: (0, import_provider_utils9.createJsonResponseHandler)(
2202
- import_zod7.z.object({
2203
- id: import_zod7.z.string(),
2204
- created_at: import_zod7.z.number(),
2205
- model: import_zod7.z.string(),
2206
- output: import_zod7.z.array(
2207
- import_zod7.z.discriminatedUnion("type", [
2208
- import_zod7.z.object({
2209
- type: import_zod7.z.literal("message"),
2210
- role: import_zod7.z.literal("assistant"),
2211
- content: import_zod7.z.array(
2212
- import_zod7.z.object({
2213
- type: import_zod7.z.literal("output_text"),
2214
- text: import_zod7.z.string(),
2215
- annotations: import_zod7.z.array(
2216
- import_zod7.z.object({
2217
- type: import_zod7.z.literal("url_citation"),
2218
- start_index: import_zod7.z.number(),
2219
- end_index: import_zod7.z.number(),
2220
- url: import_zod7.z.string(),
2221
- title: import_zod7.z.string()
2301
+ successfulResponseHandler: (0, import_provider_utils10.createJsonResponseHandler)(
2302
+ import_zod8.z.object({
2303
+ id: import_zod8.z.string(),
2304
+ created_at: import_zod8.z.number(),
2305
+ model: import_zod8.z.string(),
2306
+ output: import_zod8.z.array(
2307
+ import_zod8.z.discriminatedUnion("type", [
2308
+ import_zod8.z.object({
2309
+ type: import_zod8.z.literal("message"),
2310
+ role: import_zod8.z.literal("assistant"),
2311
+ content: import_zod8.z.array(
2312
+ import_zod8.z.object({
2313
+ type: import_zod8.z.literal("output_text"),
2314
+ text: import_zod8.z.string(),
2315
+ annotations: import_zod8.z.array(
2316
+ import_zod8.z.object({
2317
+ type: import_zod8.z.literal("url_citation"),
2318
+ start_index: import_zod8.z.number(),
2319
+ end_index: import_zod8.z.number(),
2320
+ url: import_zod8.z.string(),
2321
+ title: import_zod8.z.string()
2222
2322
  })
2223
2323
  )
2224
2324
  })
2225
2325
  )
2226
2326
  }),
2227
- import_zod7.z.object({
2228
- type: import_zod7.z.literal("function_call"),
2229
- call_id: import_zod7.z.string(),
2230
- name: import_zod7.z.string(),
2231
- arguments: import_zod7.z.string()
2327
+ import_zod8.z.object({
2328
+ type: import_zod8.z.literal("function_call"),
2329
+ call_id: import_zod8.z.string(),
2330
+ name: import_zod8.z.string(),
2331
+ arguments: import_zod8.z.string()
2232
2332
  }),
2233
- import_zod7.z.object({
2234
- type: import_zod7.z.literal("web_search_call")
2333
+ import_zod8.z.object({
2334
+ type: import_zod8.z.literal("web_search_call")
2235
2335
  }),
2236
- import_zod7.z.object({
2237
- type: import_zod7.z.literal("computer_call")
2336
+ import_zod8.z.object({
2337
+ type: import_zod8.z.literal("computer_call")
2238
2338
  }),
2239
- import_zod7.z.object({
2240
- type: import_zod7.z.literal("reasoning")
2339
+ import_zod8.z.object({
2340
+ type: import_zod8.z.literal("reasoning")
2241
2341
  })
2242
2342
  ])
2243
2343
  ),
2244
- incomplete_details: import_zod7.z.object({ reason: import_zod7.z.string() }).nullable(),
2344
+ incomplete_details: import_zod8.z.object({ reason: import_zod8.z.string() }).nullable(),
2245
2345
  usage: usageSchema
2246
2346
  })
2247
2347
  ),
@@ -2262,7 +2362,7 @@ var OpenAIResponsesLanguageModel = class {
2262
2362
  var _a2, _b2, _c2;
2263
2363
  return {
2264
2364
  sourceType: "url",
2265
- id: (_c2 = (_b2 = (_a2 = this.config).generateId) == null ? void 0 : _b2.call(_a2)) != null ? _c2 : (0, import_provider_utils9.generateId)(),
2365
+ id: (_c2 = (_b2 = (_a2 = this.config).generateId) == null ? void 0 : _b2.call(_a2)) != null ? _c2 : (0, import_provider_utils10.generateId)(),
2266
2366
  url: annotation.url,
2267
2367
  title: annotation.title
2268
2368
  };
@@ -2305,18 +2405,18 @@ var OpenAIResponsesLanguageModel = class {
2305
2405
  }
2306
2406
  async doStream(options) {
2307
2407
  const { args: body, warnings } = this.getArgs(options);
2308
- const { responseHeaders, value: response } = await (0, import_provider_utils9.postJsonToApi)({
2408
+ const { responseHeaders, value: response } = await (0, import_provider_utils10.postJsonToApi)({
2309
2409
  url: this.config.url({
2310
2410
  path: "/responses",
2311
2411
  modelId: this.modelId
2312
2412
  }),
2313
- headers: (0, import_provider_utils9.combineHeaders)(this.config.headers(), options.headers),
2413
+ headers: (0, import_provider_utils10.combineHeaders)(this.config.headers(), options.headers),
2314
2414
  body: {
2315
2415
  ...body,
2316
2416
  stream: true
2317
2417
  },
2318
2418
  failedResponseHandler: openaiFailedResponseHandler,
2319
- successfulResponseHandler: (0, import_provider_utils9.createEventSourceResponseHandler)(
2419
+ successfulResponseHandler: (0, import_provider_utils10.createEventSourceResponseHandler)(
2320
2420
  openaiResponsesChunkSchema
2321
2421
  ),
2322
2422
  abortSignal: options.abortSignal,
@@ -2404,7 +2504,7 @@ var OpenAIResponsesLanguageModel = class {
2404
2504
  type: "source",
2405
2505
  source: {
2406
2506
  sourceType: "url",
2407
- id: (_h = (_g = (_f = self.config).generateId) == null ? void 0 : _g.call(_f)) != null ? _h : (0, import_provider_utils9.generateId)(),
2507
+ id: (_h = (_g = (_f = self.config).generateId) == null ? void 0 : _g.call(_f)) != null ? _h : (0, import_provider_utils10.generateId)(),
2408
2508
  url: value.annotation.url,
2409
2509
  title: value.annotation.title
2410
2510
  }
@@ -2439,79 +2539,79 @@ var OpenAIResponsesLanguageModel = class {
2439
2539
  };
2440
2540
  }
2441
2541
  };
2442
- var usageSchema = import_zod7.z.object({
2443
- input_tokens: import_zod7.z.number(),
2444
- input_tokens_details: import_zod7.z.object({ cached_tokens: import_zod7.z.number().nullish() }).nullish(),
2445
- output_tokens: import_zod7.z.number(),
2446
- output_tokens_details: import_zod7.z.object({ reasoning_tokens: import_zod7.z.number().nullish() }).nullish()
2542
+ var usageSchema = import_zod8.z.object({
2543
+ input_tokens: import_zod8.z.number(),
2544
+ input_tokens_details: import_zod8.z.object({ cached_tokens: import_zod8.z.number().nullish() }).nullish(),
2545
+ output_tokens: import_zod8.z.number(),
2546
+ output_tokens_details: import_zod8.z.object({ reasoning_tokens: import_zod8.z.number().nullish() }).nullish()
2447
2547
  });
2448
- var textDeltaChunkSchema = import_zod7.z.object({
2449
- type: import_zod7.z.literal("response.output_text.delta"),
2450
- delta: import_zod7.z.string()
2548
+ var textDeltaChunkSchema = import_zod8.z.object({
2549
+ type: import_zod8.z.literal("response.output_text.delta"),
2550
+ delta: import_zod8.z.string()
2451
2551
  });
2452
- var responseFinishedChunkSchema = import_zod7.z.object({
2453
- type: import_zod7.z.enum(["response.completed", "response.incomplete"]),
2454
- response: import_zod7.z.object({
2455
- incomplete_details: import_zod7.z.object({ reason: import_zod7.z.string() }).nullish(),
2552
+ var responseFinishedChunkSchema = import_zod8.z.object({
2553
+ type: import_zod8.z.enum(["response.completed", "response.incomplete"]),
2554
+ response: import_zod8.z.object({
2555
+ incomplete_details: import_zod8.z.object({ reason: import_zod8.z.string() }).nullish(),
2456
2556
  usage: usageSchema
2457
2557
  })
2458
2558
  });
2459
- var responseCreatedChunkSchema = import_zod7.z.object({
2460
- type: import_zod7.z.literal("response.created"),
2461
- response: import_zod7.z.object({
2462
- id: import_zod7.z.string(),
2463
- created_at: import_zod7.z.number(),
2464
- model: import_zod7.z.string()
2559
+ var responseCreatedChunkSchema = import_zod8.z.object({
2560
+ type: import_zod8.z.literal("response.created"),
2561
+ response: import_zod8.z.object({
2562
+ id: import_zod8.z.string(),
2563
+ created_at: import_zod8.z.number(),
2564
+ model: import_zod8.z.string()
2465
2565
  })
2466
2566
  });
2467
- var responseOutputItemDoneSchema = import_zod7.z.object({
2468
- type: import_zod7.z.literal("response.output_item.done"),
2469
- output_index: import_zod7.z.number(),
2470
- item: import_zod7.z.discriminatedUnion("type", [
2471
- import_zod7.z.object({
2472
- type: import_zod7.z.literal("message")
2567
+ var responseOutputItemDoneSchema = import_zod8.z.object({
2568
+ type: import_zod8.z.literal("response.output_item.done"),
2569
+ output_index: import_zod8.z.number(),
2570
+ item: import_zod8.z.discriminatedUnion("type", [
2571
+ import_zod8.z.object({
2572
+ type: import_zod8.z.literal("message")
2473
2573
  }),
2474
- import_zod7.z.object({
2475
- type: import_zod7.z.literal("function_call"),
2476
- id: import_zod7.z.string(),
2477
- call_id: import_zod7.z.string(),
2478
- name: import_zod7.z.string(),
2479
- arguments: import_zod7.z.string(),
2480
- status: import_zod7.z.literal("completed")
2574
+ import_zod8.z.object({
2575
+ type: import_zod8.z.literal("function_call"),
2576
+ id: import_zod8.z.string(),
2577
+ call_id: import_zod8.z.string(),
2578
+ name: import_zod8.z.string(),
2579
+ arguments: import_zod8.z.string(),
2580
+ status: import_zod8.z.literal("completed")
2481
2581
  })
2482
2582
  ])
2483
2583
  });
2484
- var responseFunctionCallArgumentsDeltaSchema = import_zod7.z.object({
2485
- type: import_zod7.z.literal("response.function_call_arguments.delta"),
2486
- item_id: import_zod7.z.string(),
2487
- output_index: import_zod7.z.number(),
2488
- delta: import_zod7.z.string()
2584
+ var responseFunctionCallArgumentsDeltaSchema = import_zod8.z.object({
2585
+ type: import_zod8.z.literal("response.function_call_arguments.delta"),
2586
+ item_id: import_zod8.z.string(),
2587
+ output_index: import_zod8.z.number(),
2588
+ delta: import_zod8.z.string()
2489
2589
  });
2490
- var responseOutputItemAddedSchema = import_zod7.z.object({
2491
- type: import_zod7.z.literal("response.output_item.added"),
2492
- output_index: import_zod7.z.number(),
2493
- item: import_zod7.z.discriminatedUnion("type", [
2494
- import_zod7.z.object({
2495
- type: import_zod7.z.literal("message")
2590
+ var responseOutputItemAddedSchema = import_zod8.z.object({
2591
+ type: import_zod8.z.literal("response.output_item.added"),
2592
+ output_index: import_zod8.z.number(),
2593
+ item: import_zod8.z.discriminatedUnion("type", [
2594
+ import_zod8.z.object({
2595
+ type: import_zod8.z.literal("message")
2496
2596
  }),
2497
- import_zod7.z.object({
2498
- type: import_zod7.z.literal("function_call"),
2499
- id: import_zod7.z.string(),
2500
- call_id: import_zod7.z.string(),
2501
- name: import_zod7.z.string(),
2502
- arguments: import_zod7.z.string()
2597
+ import_zod8.z.object({
2598
+ type: import_zod8.z.literal("function_call"),
2599
+ id: import_zod8.z.string(),
2600
+ call_id: import_zod8.z.string(),
2601
+ name: import_zod8.z.string(),
2602
+ arguments: import_zod8.z.string()
2503
2603
  })
2504
2604
  ])
2505
2605
  });
2506
- var responseAnnotationAddedSchema = import_zod7.z.object({
2507
- type: import_zod7.z.literal("response.output_text.annotation.added"),
2508
- annotation: import_zod7.z.object({
2509
- type: import_zod7.z.literal("url_citation"),
2510
- url: import_zod7.z.string(),
2511
- title: import_zod7.z.string()
2606
+ var responseAnnotationAddedSchema = import_zod8.z.object({
2607
+ type: import_zod8.z.literal("response.output_text.annotation.added"),
2608
+ annotation: import_zod8.z.object({
2609
+ type: import_zod8.z.literal("url_citation"),
2610
+ url: import_zod8.z.string(),
2611
+ title: import_zod8.z.string()
2512
2612
  })
2513
2613
  });
2514
- var openaiResponsesChunkSchema = import_zod7.z.union([
2614
+ var openaiResponsesChunkSchema = import_zod8.z.union([
2515
2615
  textDeltaChunkSchema,
2516
2616
  responseFinishedChunkSchema,
2517
2617
  responseCreatedChunkSchema,
@@ -2519,7 +2619,7 @@ var openaiResponsesChunkSchema = import_zod7.z.union([
2519
2619
  responseFunctionCallArgumentsDeltaSchema,
2520
2620
  responseOutputItemAddedSchema,
2521
2621
  responseAnnotationAddedSchema,
2522
- import_zod7.z.object({ type: import_zod7.z.string() }).passthrough()
2622
+ import_zod8.z.object({ type: import_zod8.z.string() }).passthrough()
2523
2623
  // fallback for unknown chunks
2524
2624
  ]);
2525
2625
  function isTextDeltaChunk(chunk) {
@@ -2564,15 +2664,15 @@ function getResponsesModelConfig(modelId) {
2564
2664
  requiredAutoTruncation: false
2565
2665
  };
2566
2666
  }
2567
- var openaiResponsesProviderOptionsSchema = import_zod7.z.object({
2568
- metadata: import_zod7.z.any().nullish(),
2569
- parallelToolCalls: import_zod7.z.boolean().nullish(),
2570
- previousResponseId: import_zod7.z.string().nullish(),
2571
- store: import_zod7.z.boolean().nullish(),
2572
- user: import_zod7.z.string().nullish(),
2573
- reasoningEffort: import_zod7.z.string().nullish(),
2574
- strictSchemas: import_zod7.z.boolean().nullish(),
2575
- instructions: import_zod7.z.string().nullish()
2667
+ var openaiResponsesProviderOptionsSchema = import_zod8.z.object({
2668
+ metadata: import_zod8.z.any().nullish(),
2669
+ parallelToolCalls: import_zod8.z.boolean().nullish(),
2670
+ previousResponseId: import_zod8.z.string().nullish(),
2671
+ store: import_zod8.z.boolean().nullish(),
2672
+ user: import_zod8.z.string().nullish(),
2673
+ reasoningEffort: import_zod8.z.string().nullish(),
2674
+ strictSchemas: import_zod8.z.boolean().nullish(),
2675
+ instructions: import_zod8.z.string().nullish()
2576
2676
  });
2577
2677
  // Annotate the CommonJS export names for ESM import in node:
2578
2678
  0 && (module.exports = {
@@ -2581,6 +2681,7 @@ var openaiResponsesProviderOptionsSchema = import_zod7.z.object({
2581
2681
  OpenAIEmbeddingModel,
2582
2682
  OpenAIImageModel,
2583
2683
  OpenAIResponsesLanguageModel,
2684
+ OpenAISpeechModel,
2584
2685
  OpenAITranscriptionModel,
2585
2686
  modelMaxImagesPerCall
2586
2687
  });