genkitx-openai - npm Package Compare versions

Comparing version 0.20.2 to 0.20.3

lib/embedder.d.ts

 import 'genkit';
-export { S as SUPPORTED_EMBEDDING_MODELS, T as TextEmbeddingConfigSchema, c as TextEmbeddingGeckoConfig, d as TextEmbeddingInputSchema, e as openaiEmbedder, t as textEmbedding3Large, a as textEmbedding3Small, b as textEmbeddingAda002 } from './embedder-DXIcCYcu.js';
+export { S as SUPPORTED_EMBEDDING_MODELS, T as TextEmbeddingConfigSchema, c as TextEmbeddingGeckoConfig, d as TextEmbeddingInputSchema, e as openaiEmbedder, t as textEmbedding3Large, a as textEmbedding3Small, b as textEmbeddingAda002 } from './embedder-DIOltsXz.js';
 import 'genkit/plugin';
 import 'openai';
+import 'genkit/model';
 import './dalle.js';
-import 'genkit/model';
 import './gpt.js';
@@ -8,0 +8,0 @@ import 'openai/resources/index.mjs';

lib/gpt.d.ts

@@ -305,2 +305,49 @@ import { CandidateData, GenerateResponseChunkData, ModelInfo, ModelAction } from 'genkit/model';
 }>>;
+declare const o3: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    frequencyPenalty: z.ZodOptional<z.ZodNumber>;
+    logitBias: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodNumber>>;
+    logProbs: z.ZodOptional<z.ZodBoolean>;
+    presencePenalty: z.ZodOptional<z.ZodNumber>;
+    seed: z.ZodOptional<z.ZodNumber>;
+    topLogProbs: z.ZodOptional<z.ZodNumber>;
+    user: z.ZodOptional<z.ZodString>;
+    visualDetailLevel: z.ZodOptional<z.ZodEnum<["auto", "low", "high"]>>;
+}>, "strip", z.ZodTypeAny, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}>>;
 declare const o3Mini: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
@@ -353,2 +400,49 @@ version: z.ZodOptional<z.ZodString>;
 }>>;
+declare const o4Mini: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    frequencyPenalty: z.ZodOptional<z.ZodNumber>;
+    logitBias: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodNumber>>;
+    logProbs: z.ZodOptional<z.ZodBoolean>;
+    presencePenalty: z.ZodOptional<z.ZodNumber>;
+    seed: z.ZodOptional<z.ZodNumber>;
+    topLogProbs: z.ZodOptional<z.ZodNumber>;
+    user: z.ZodOptional<z.ZodString>;
+    visualDetailLevel: z.ZodOptional<z.ZodEnum<["auto", "low", "high"]>>;
+}>, "strip", z.ZodTypeAny, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}>>;
 declare const gpt4oMini: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
@@ -542,2 +636,143 @@ version: z.ZodOptional<z.ZodString>;
 }>>;
+declare const gpt41: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    frequencyPenalty: z.ZodOptional<z.ZodNumber>;
+    logitBias: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodNumber>>;
+    logProbs: z.ZodOptional<z.ZodBoolean>;
+    presencePenalty: z.ZodOptional<z.ZodNumber>;
+    seed: z.ZodOptional<z.ZodNumber>;
+    topLogProbs: z.ZodOptional<z.ZodNumber>;
+    user: z.ZodOptional<z.ZodString>;
+    visualDetailLevel: z.ZodOptional<z.ZodEnum<["auto", "low", "high"]>>;
+}>, "strip", z.ZodTypeAny, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}>>;
+declare const gpt41Mini: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    frequencyPenalty: z.ZodOptional<z.ZodNumber>;
+    logitBias: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodNumber>>;
+    logProbs: z.ZodOptional<z.ZodBoolean>;
+    presencePenalty: z.ZodOptional<z.ZodNumber>;
+    seed: z.ZodOptional<z.ZodNumber>;
+    topLogProbs: z.ZodOptional<z.ZodNumber>;
+    user: z.ZodOptional<z.ZodString>;
+    visualDetailLevel: z.ZodOptional<z.ZodEnum<["auto", "low", "high"]>>;
+}>, "strip", z.ZodTypeAny, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}>>;
+declare const gpt41Nano: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    frequencyPenalty: z.ZodOptional<z.ZodNumber>;
+    logitBias: z.ZodOptional<z.ZodRecord<z.ZodString, z.ZodNumber>>;
+    logProbs: z.ZodOptional<z.ZodBoolean>;
+    presencePenalty: z.ZodOptional<z.ZodNumber>;
+    seed: z.ZodOptional<z.ZodNumber>;
+    topLogProbs: z.ZodOptional<z.ZodNumber>;
+    user: z.ZodOptional<z.ZodString>;
+    visualDetailLevel: z.ZodOptional<z.ZodEnum<["auto", "low", "high"]>>;
+}>, "strip", z.ZodTypeAny, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}, {
+    user?: string | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    frequencyPenalty?: number | undefined;
+    logitBias?: Record<string, number> | undefined;
+    logProbs?: boolean | undefined;
+    presencePenalty?: number | undefined;
+    seed?: number | undefined;
+    topLogProbs?: number | undefined;
+    visualDetailLevel?: "auto" | "low" | "high" | undefined;
+}>>;
 declare const gpt35Turbo: ModelReference<z.ZodObject<z.objectUtil.extendShape<{
@@ -651,2 +886,2 @@ version: z.ZodOptional<z.ZodString>;
 
-export { OpenAiConfigSchema, SUPPORTED_GPT_MODELS, fromOpenAiChoice, fromOpenAiChunkChoice, fromOpenAiToolCall, gpt35Turbo, gpt4, gpt45, gpt4Turbo, gpt4Vision, gpt4o, gpt4oMini, gptModel, gptRunner, o1, o1Mini, o1Preview, o3Mini, toOpenAIRole, toOpenAiMessages, toOpenAiRequestBody, toOpenAiTextAndMedia };
+export { OpenAiConfigSchema, SUPPORTED_GPT_MODELS, fromOpenAiChoice, fromOpenAiChunkChoice, fromOpenAiToolCall, gpt35Turbo, gpt4, gpt41, gpt41Mini, gpt41Nano, gpt45, gpt4Turbo, gpt4Vision, gpt4o, gpt4oMini, gptModel, gptRunner, o1, o1Mini, o1Preview, o3, o3Mini, o4Mini, toOpenAIRole, toOpenAiMessages, toOpenAiRequestBody, toOpenAiTextAndMedia };

lib/gpt.js

@@ -67,2 +67,5 @@ "use strict";
   gpt4: () => gpt4,
+  gpt41: () => gpt41,
+  gpt41Mini: () => gpt41Mini,
+  gpt41Nano: () => gpt41Nano,
   gpt45: () => gpt45,
@@ -78,3 +81,5 @@ gpt4Turbo: () => gpt4Turbo,
   o1Preview: () => o1Preview,
+  o3: () => o3,
   o3Mini: () => o3Mini,
+  o4Mini: () => o4Mini,
   toOpenAIRole: () => toOpenAIRole,
@@ -189,2 +194,17 @@ toOpenAiMessages: () => toOpenAiMessages,
 });
+const o3 = (0, import_model.modelRef)({
+  name: "openai/o3",
+  info: {
+    versions: ["o3"],
+    label: "OpenAI - o3",
+    supports: {
+      multiturn: true,
+      tools: true,
+      media: false,
+      systemRole: false,
+      output: ["text", "json"]
+    }
+  },
+  configSchema: OpenAiConfigSchema
+});
 const o3Mini = (0, import_model.modelRef)({
@@ -205,2 +225,17 @@ name: "openai/o3-mini",
 });
+const o4Mini = (0, import_model.modelRef)({
+  name: "openai/o4-mini",
+  info: {
+    versions: ["o4-mini"],
+    label: "OpenAI - o4 Mini",
+    supports: {
+      multiturn: true,
+      tools: true,
+      media: false,
+      systemRole: false,
+      output: ["text", "json"]
+    }
+  },
+  configSchema: OpenAiConfigSchema
+});
 const gpt4oMini = (0, import_model.modelRef)({
@@ -272,2 +307,47 @@ name: "openai/gpt-4o-mini",
 });
+const gpt41 = (0, import_model.modelRef)({
+  name: "openai/gpt-4.1",
+  info: {
+    versions: ["gpt-4.1"],
+    label: "OpenAI - GPT-4.1",
+    supports: {
+      multiturn: true,
+      tools: true,
+      media: false,
+      systemRole: true,
+      output: ["text", "json"]
+    }
+  },
+  configSchema: OpenAiConfigSchema
+});
+const gpt41Mini = (0, import_model.modelRef)({
+  name: "openai/gpt-4.1-mini",
+  info: {
+    versions: ["gpt-4.1-mini"],
+    label: "OpenAI - GPT-4.1 Mini",
+    supports: {
+      multiturn: true,
+      tools: true,
+      media: false,
+      systemRole: true,
+      output: ["text", "json"]
+    }
+  },
+  configSchema: OpenAiConfigSchema
+});
+const gpt41Nano = (0, import_model.modelRef)({
+  name: "openai/gpt-4.1-nano",
+  info: {
+    versions: ["gpt-4.1-nano"],
+    label: "OpenAI - GPT-4.1 Nano",
+    supports: {
+      multiturn: true,
+      tools: true,
+      media: false,
+      systemRole: true,
+      output: ["text", "json"]
+    }
+  },
+  configSchema: OpenAiConfigSchema
+});
 const gpt35Turbo = (0, import_model.modelRef)({
@@ -295,2 +375,5 @@ name: "openai/gpt-3.5-turbo",
   "gpt-4": gpt4,
+  "gpt-4.1": gpt41,
+  "gpt-4.1-mini": gpt41Mini,
+  "gpt-4.1-nano": gpt41Nano,
   "gpt-3.5-turbo": gpt35Turbo,
@@ -300,3 +383,5 @@ "o1-preview": o1Preview,
   "o1-mini": o1Mini,
-  "o3-mini": o3Mini
+  o3,
+  "o3-mini": o3Mini,
+  "o4-mini": o4Mini
 };
@@ -624,2 +709,5 @@ function toOpenAIRole(role) {
   gpt4,
+  gpt41,
+  gpt41Mini,
+  gpt41Nano,
   gpt45,
@@ -635,3 +723,5 @@ gpt4Turbo,
   o1Preview,
+  o3,
   o3Mini,
+  o4Mini,
   toOpenAIRole,
@@ -638,0 +728,0 @@ toOpenAiMessages,

lib/index.d.ts

 import 'genkit/plugin';
 import 'openai';
+import 'genkit/model';
 export { dallE3 } from './dalle.js';
-export { M as ModelDefinition, P as PluginOptions, o as default, o as openAI, t as textEmbedding3Large, a as textEmbedding3Small, b as textEmbeddingAda002 } from './embedder-DXIcCYcu.js';
-export { gpt35Turbo, gpt4, gpt45, gpt4Turbo, gpt4Vision, gpt4o, gpt4oMini, o1, o1Mini, o1Preview, o3Mini } from './gpt.js';
-export { tts1, tts1Hd } from './tts.js';
-export { whisper1 } from './whisper.js';
-import 'genkit/model';
+export { M as ModelDefinition, P as PluginOptions, o as default, o as openAI, t as textEmbedding3Large, a as textEmbedding3Small, b as textEmbeddingAda002 } from './embedder-DIOltsXz.js';
+export { gpt35Turbo, gpt4, gpt41, gpt41Mini, gpt41Nano, gpt45, gpt4Turbo, gpt4Vision, gpt4o, gpt4oMini, o1, o1Mini, o1Preview, o3, o3Mini, o4Mini } from './gpt.js';
+export { gpt4oMiniTts, tts1, tts1Hd } from './tts.js';
+export { gpt4oTranscribe, whisper1 } from './whisper.js';
 import 'genkit';
 import 'openai/resources/index.mjs';

lib/index.js

@@ -45,2 +45,5 @@ "use strict";
   gpt4: () => import_gpt.gpt4,
+  gpt41: () => import_gpt.gpt41,
+  gpt41Mini: () => import_gpt.gpt41Mini,
+  gpt41Nano: () => import_gpt.gpt41Nano,
   gpt45: () => import_gpt.gpt45,
@@ -51,6 +54,10 @@ gpt4Turbo: () => import_gpt.gpt4Turbo,
   gpt4oMini: () => import_gpt.gpt4oMini,
+  gpt4oMiniTts: () => import_tts.gpt4oMiniTts,
+  gpt4oTranscribe: () => import_whisper.gpt4oTranscribe,
   o1: () => import_gpt.o1,
   o1Mini: () => import_gpt.o1Mini,
   o1Preview: () => import_gpt.o1Preview,
+  o3: () => import_gpt.o3,
   o3Mini: () => import_gpt.o3Mini,
+  o4Mini: () => import_gpt.o4Mini,
   openAI: () => openAI,
@@ -85,3 +92,5 @@ textEmbedding3Large: () => import_embedder.textEmbedding3Large,
   (0, import_dalle.dallE3Model)(ai, client);
-  (0, import_whisper.whisper1Model)(ai, client);
+  for (const name of Object.keys(import_whisper.SUPPORTED_STT_MODELS)) {
+    (0, import_whisper.sttModel)(ai, name, client);
+  }
   for (const name of Object.keys(import_tts.SUPPORTED_TTS_MODELS)) {
@@ -100,2 +109,5 @@ (0, import_tts.ttsModel)(ai, name, client);
   gpt4,
+  gpt41,
+  gpt41Mini,
+  gpt41Nano,
   gpt45,
@@ -106,6 +118,10 @@ gpt4Turbo,
   gpt4oMini,
+  gpt4oMiniTts,
+  gpt4oTranscribe,
   o1,
   o1Mini,
   o1Preview,
+  o3,
   o3Mini,
+  o4Mini,
   openAI,
@@ -112,0 +128,0 @@ textEmbedding3Large,

lib/tts.d.ts

@@ -102,2 +102,34 @@ import * as genkit from 'genkit';
 }>>;
+declare const gpt4oMiniTts: genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    voice: z.ZodDefault<z.ZodOptional<z.ZodEnum<["alloy", "echo", "fable", "onyx", "nova", "shimmer"]>>>;
+    speed: z.ZodOptional<z.ZodNumber>;
+    response_format: z.ZodOptional<z.ZodEnum<["mp3", "opus", "aac", "flac", "wav", "pcm"]>>;
+}>, "strip", z.ZodTypeAny, {
+    voice: "alloy" | "echo" | "fable" | "onyx" | "nova" | "shimmer";
+    response_format?: "mp3" | "opus" | "aac" | "flac" | "wav" | "pcm" | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    speed?: number | undefined;
+}, {
+    response_format?: "mp3" | "opus" | "aac" | "flac" | "wav" | "pcm" | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    voice?: "alloy" | "echo" | "fable" | "onyx" | "nova" | "shimmer" | undefined;
+    speed?: number | undefined;
+}>>;
 declare const SUPPORTED_TTS_MODELS: {
@@ -168,2 +200,34 @@ 'tts-1': genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
     }>>;
+    'gpt-4o-mini-tts': genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+        version: z.ZodOptional<z.ZodString>;
+        temperature: z.ZodOptional<z.ZodNumber>;
+        maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+        topK: z.ZodOptional<z.ZodNumber>;
+        topP: z.ZodOptional<z.ZodNumber>;
+        stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+    }, {
+        voice: z.ZodDefault<z.ZodOptional<z.ZodEnum<["alloy", "echo", "fable", "onyx", "nova", "shimmer"]>>>;
+        speed: z.ZodOptional<z.ZodNumber>;
+        response_format: z.ZodOptional<z.ZodEnum<["mp3", "opus", "aac", "flac", "wav", "pcm"]>>;
+    }>, "strip", z.ZodTypeAny, {
+        voice: "alloy" | "echo" | "fable" | "onyx" | "nova" | "shimmer";
+        response_format?: "mp3" | "opus" | "aac" | "flac" | "wav" | "pcm" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        speed?: number | undefined;
+    }, {
+        response_format?: "mp3" | "opus" | "aac" | "flac" | "wav" | "pcm" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        voice?: "alloy" | "echo" | "fable" | "onyx" | "nova" | "shimmer" | undefined;
+        speed?: number | undefined;
+    }>>;
 };
@@ -180,2 +244,2 @@ declare const RESPONSE_FORMAT_MEDIA_TYPES: {
 
-export { RESPONSE_FORMAT_MEDIA_TYPES, SUPPORTED_TTS_MODELS, TTSConfigSchema, tts1, tts1Hd, ttsModel };
+export { RESPONSE_FORMAT_MEDIA_TYPES, SUPPORTED_TTS_MODELS, TTSConfigSchema, gpt4oMiniTts, tts1, tts1Hd, ttsModel };

lib/tts.js

@@ -61,2 +61,3 @@ "use strict";
   TTSConfigSchema: () => TTSConfigSchema,
+  gpt4oMiniTts: () => gpt4oMiniTts,
   tts1: () => tts1,
@@ -102,5 +103,20 @@ tts1Hd: () => tts1Hd,
 });
+const gpt4oMiniTts = (0, import_model.modelRef)({
+  name: "openai/gpt-4o-mini-tts",
+  info: {
+    label: "OpenAI - GPT-4o Mini Text-to-speech",
+    supports: {
+      media: false,
+      output: ["media"],
+      multiturn: false,
+      systemRole: false,
+      tools: false
+    }
+  },
+  configSchema: TTSConfigSchema
+});
 const SUPPORTED_TTS_MODELS = {
   "tts-1": tts1,
-  "tts-1-hd": tts1Hd
+  "tts-1-hd": tts1Hd,
+  "gpt-4o-mini-tts": gpt4oMiniTts
 };
@@ -178,2 +194,3 @@ const RESPONSE_FORMAT_MEDIA_TYPES = {
   TTSConfigSchema,
+  gpt4oMiniTts,
   tts1,
@@ -180,0 +197,0 @@ tts1Hd,

lib/whisper.d.ts

@@ -70,4 +70,102 @@ import * as genkit from 'genkit';
 }>>;
-declare function whisper1Model(ai: Genkit, client: OpenAI): ModelAction<typeof Whisper1ConfigSchema>;
+declare const gpt4oTranscribe: genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+    version: z.ZodOptional<z.ZodString>;
+    temperature: z.ZodOptional<z.ZodNumber>;
+    maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+    topK: z.ZodOptional<z.ZodNumber>;
+    topP: z.ZodOptional<z.ZodNumber>;
+    stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+}, {
+    language: z.ZodOptional<z.ZodString>;
+    timestamp_granularities: z.ZodOptional<z.ZodArray<z.ZodEnum<["word", "segment"]>, "many">>;
+    response_format: z.ZodOptional<z.ZodEnum<["json", "text", "srt", "verbose_json", "vtt"]>>;
+}>, "strip", z.ZodTypeAny, {
+    response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    language?: string | undefined;
+    timestamp_granularities?: ("word" | "segment")[] | undefined;
+}, {
+    response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+    version?: string | undefined;
+    temperature?: number | undefined;
+    maxOutputTokens?: number | undefined;
+    topK?: number | undefined;
+    topP?: number | undefined;
+    stopSequences?: string[] | undefined;
+    language?: string | undefined;
+    timestamp_granularities?: ("word" | "segment")[] | undefined;
+}>>;
+declare const SUPPORTED_STT_MODELS: {
+    'gpt-4o-transcribe': genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+        version: z.ZodOptional<z.ZodString>;
+        temperature: z.ZodOptional<z.ZodNumber>;
+        maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+        topK: z.ZodOptional<z.ZodNumber>;
+        topP: z.ZodOptional<z.ZodNumber>;
+        stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+    }, {
+        language: z.ZodOptional<z.ZodString>;
+        timestamp_granularities: z.ZodOptional<z.ZodArray<z.ZodEnum<["word", "segment"]>, "many">>;
+        response_format: z.ZodOptional<z.ZodEnum<["json", "text", "srt", "verbose_json", "vtt"]>>;
+    }>, "strip", z.ZodTypeAny, {
+        response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        language?: string | undefined;
+        timestamp_granularities?: ("word" | "segment")[] | undefined;
+    }, {
+        response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        language?: string | undefined;
+        timestamp_granularities?: ("word" | "segment")[] | undefined;
+    }>>;
+    'whisper-1': genkit.ModelReference<z.ZodObject<z.objectUtil.extendShape<{
+        version: z.ZodOptional<z.ZodString>;
+        temperature: z.ZodOptional<z.ZodNumber>;
+        maxOutputTokens: z.ZodOptional<z.ZodNumber>;
+        topK: z.ZodOptional<z.ZodNumber>;
+        topP: z.ZodOptional<z.ZodNumber>;
+        stopSequences: z.ZodOptional<z.ZodArray<z.ZodString, "many">>;
+    }, {
+        language: z.ZodOptional<z.ZodString>;
+        timestamp_granularities: z.ZodOptional<z.ZodArray<z.ZodEnum<["word", "segment"]>, "many">>;
+        response_format: z.ZodOptional<z.ZodEnum<["json", "text", "srt", "verbose_json", "vtt"]>>;
+    }>, "strip", z.ZodTypeAny, {
+        response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        language?: string | undefined;
+        timestamp_granularities?: ("word" | "segment")[] | undefined;
+    }, {
+        response_format?: "text" | "json" | "srt" | "verbose_json" | "vtt" | undefined;
+        version?: string | undefined;
+        temperature?: number | undefined;
+        maxOutputTokens?: number | undefined;
+        topK?: number | undefined;
+        topP?: number | undefined;
+        stopSequences?: string[] | undefined;
+        language?: string | undefined;
+        timestamp_granularities?: ("word" | "segment")[] | undefined;
+    }>>;
+};
+declare function sttModel(ai: Genkit, name: string, client: OpenAI): ModelAction<typeof Whisper1ConfigSchema>;
 
-export { Whisper1ConfigSchema, whisper1, whisper1Model };
+export { SUPPORTED_STT_MODELS, Whisper1ConfigSchema, gpt4oTranscribe, sttModel, whisper1 };

lib/whisper.js

@@ -58,5 +58,7 @@ "use strict";
 __export(whisper_exports, {
+  SUPPORTED_STT_MODELS: () => SUPPORTED_STT_MODELS,
   Whisper1ConfigSchema: () => Whisper1ConfigSchema,
-  whisper1: () => whisper1,
-  whisper1Model: () => whisper1Model
+  gpt4oTranscribe: () => gpt4oTranscribe,
+  sttModel: () => sttModel,
+  whisper1: () => whisper1
 });
@@ -85,2 +87,16 @@ module.exports = __toCommonJS(whisper_exports);
 });
+const gpt4oTranscribe = (0, import_model.modelRef)({
+  name: "openai/gpt-4o-transcribe",
+  info: {
+    label: "OpenAI - GPT-4o Transcribe",
+    supports: {
+      media: true,
+      output: ["text", "json"],
+      multiturn: false,
+      systemRole: false,
+      tools: false
+    }
+  },
+  configSchema: Whisper1ConfigSchema
+});
 function toWhisper1Request(request) {
@@ -146,13 +162,21 @@ var _a, _b, _c, _d, _e, _f;
 }
-function whisper1Model(ai, client) {
+const SUPPORTED_STT_MODELS = {
+  "gpt-4o-transcribe": gpt4oTranscribe,
+  "whisper-1": whisper1
+};
+function sttModel(ai, name, client) {
+  const modelId = `openai/${name}`;
+  const model = SUPPORTED_STT_MODELS[name];
+  if (!model) throw new Error(`Unsupported model: ${name}`);
   return ai.defineModel(
     __spreadProps(__spreadValues({
-      name: whisper1.name
-    }, whisper1.info), {
-      configSchema: whisper1.configSchema
+      name: modelId
+    }, model.info), {
+      configSchema: model.configSchema
     }),
     (request) => __async(this, null, function* () {
-      const result = yield client.audio.transcriptions.create(
-        toWhisper1Request(request)
-      );
+      const params = toWhisper1Request(request);
+      const result = yield client.audio.transcriptions.create(__spreadProps(__spreadValues({}, params), {
+        stream: false
+      }));
       return toGenerateResponse(result);
@@ -164,6 +188,8 @@ })
 0 && (module.exports = {
+  SUPPORTED_STT_MODELS,
   Whisper1ConfigSchema,
-  whisper1,
-  whisper1Model
+  gpt4oTranscribe,
+  sttModel,
+  whisper1
 });
 //# sourceMappingURL=whisper.js.map

package.json

@@ -17,3 +17,3 @@ {
   ],
-  "version": "0.20.2",
+  "version": "0.20.3",
   "type": "commonjs",
@@ -28,3 +28,3 @@ "repository": {
   "dependencies": {
-    "openai": "^4.47.1"
+    "openai": "^4.95.0"
   },
@@ -67,3 +67,3 @@ "peerDependencies": {
   },
-  "gitHead": "07d893c51ed1b64d2a33c7be7c6ad494bbe9a01d"
+  "gitHead": "6dcb587c5aa87b090c3313590f690521c9375f19"
 }

No alert changes

Improved metrics

Total package byte prevSize

361555

16.74%

Lines of code

4282

19.41%

Dependency changes

• Updated

  openai@^4.95.0