ai 6.0.0-beta.103 → 6.0.0-beta.104

package/CHANGELOG.md

@@ -1,5 +1,13 @@
 # ai
 
+## 6.0.0-beta.104
+
+### Patch Changes
+
+- 2d166e4: feat(provider/gateway): add support for image models
+- Updated dependencies [2d166e4]
+  - @ai-sdk/gateway@2.0.0-beta.56
+
 ## 6.0.0-beta.103
 
 ### Patch Changes

package/dist/index.d.mts

@@ -3,7 +3,7 @@ import * as _ai_sdk_provider_utils from '@ai-sdk/provider-utils';
 import { Tool, InferToolInput, InferToolOutput, FlexibleSchema, InferSchema, AssistantModelMessage, ToolModelMessage, ReasoningPart, ModelMessage, SystemModelMessage, UserModelMessage, ProviderOptions, IdGenerator, ToolCall, MaybePromiseLike, TextPart, FilePart, Resolvable, FetchFunction, DataContent } from '@ai-sdk/provider-utils';
 export { AssistantContent, AssistantModelMessage, DataContent, FilePart, FlexibleSchema, IdGenerator, ImagePart, InferSchema, InferToolInput, InferToolOutput, ModelMessage, Schema, SystemModelMessage, TextPart, Tool, ToolApprovalRequest, ToolApprovalResponse, ToolCallOptions, ToolCallPart, ToolContent, ToolExecuteFunction, ToolModelMessage, ToolResultPart, UserContent, UserModelMessage, asSchema, createIdGenerator, dynamicTool, generateId, jsonSchema, parseJsonEventStream, tool, zodSchema } from '@ai-sdk/provider-utils';
 import * as _ai_sdk_provider from '@ai-sdk/provider';
-import { EmbeddingModelV3, EmbeddingModelV2, EmbeddingModelV3Embedding, ImageModelV3, ImageModelV3CallWarning, ImageModelV3ProviderMetadata, JSONValue as JSONValue$1, LanguageModelV3, LanguageModelV2, LanguageModelV3FinishReason, LanguageModelV3CallWarning, LanguageModelV3Source, LanguageModelV3Middleware, EmbeddingModelV3Middleware, RerankingModelV3, SharedV3ProviderMetadata, SpeechModelV3, SpeechModelV2, SpeechModelV3CallWarning, TranscriptionModelV3, TranscriptionModelV2, TranscriptionModelV3CallWarning, LanguageModelV3Usage, ImageModelV3Usage, LanguageModelV3CallOptions, AISDKError, LanguageModelV3ToolCall, JSONSchema7, JSONParseError, TypeValidationError, JSONObject, SharedV3Warning, EmbeddingModelCallOptions, ProviderV3, ProviderV2, NoSuchModelError } from '@ai-sdk/provider';
+import { EmbeddingModelV3, EmbeddingModelV2, EmbeddingModelV3Embedding, ImageModelV3, ImageModelV2, ImageModelV3CallWarning, ImageModelV2CallWarning, ImageModelV3ProviderMetadata, ImageModelV2ProviderMetadata, JSONValue as JSONValue$1, LanguageModelV3, LanguageModelV2, LanguageModelV3FinishReason, LanguageModelV3CallWarning, LanguageModelV3Source, LanguageModelV3Middleware, EmbeddingModelV3Middleware, RerankingModelV3, SharedV3ProviderMetadata, SpeechModelV3, SpeechModelV2, SpeechModelV3CallWarning, TranscriptionModelV3, TranscriptionModelV2, TranscriptionModelV3CallWarning, LanguageModelV3Usage, ImageModelV3Usage, LanguageModelV3CallOptions, AISDKError, LanguageModelV3ToolCall, JSONSchema7, JSONParseError, TypeValidationError, JSONObject, SharedV3Warning, EmbeddingModelCallOptions, ProviderV3, ProviderV2, NoSuchModelError } from '@ai-sdk/provider';
 export { AISDKError, APICallError, EmptyResponseBodyError, InvalidPromptError, InvalidResponseDataError, JSONParseError, JSONSchema7, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, TooManyEmbeddingValuesForCallError, TypeValidationError, UnsupportedFunctionalityError } from '@ai-sdk/provider';
 import { ServerResponse } from 'node:http';
 import { AttributeValue, Tracer } from '@opentelemetry/api';
@@ -22,16 +22,16 @@ type Embedding = EmbeddingModelV3Embedding;
 /**
 Image model that is used by the AI SDK Core functions.
   */
-type ImageModel = ImageModelV3;
+type ImageModel = string | ImageModelV3 | ImageModelV2;
 /**
 Warning from the model provider for this call. The call will proceed, but e.g.
 some settings might not be supported, which can lead to suboptimal results.
   */
-type ImageGenerationWarning = ImageModelV3CallWarning;
+type ImageGenerationWarning = ImageModelV3CallWarning | ImageModelV2CallWarning;
 /**
 Metadata from the model provider for this call
   */
-type ImageModelProviderMetadata = ImageModelV3ProviderMetadata;
+type ImageModelProviderMetadata = ImageModelV3ProviderMetadata | ImageModelV2ProviderMetadata;
 
 type ImageModelResponseMetadata = {
     /**
@@ -3951,11 +3951,11 @@ as body parameters.
 
 @returns A result object that contains the generated images.
  */
-declare function generateImage({ model, prompt, n, maxImagesPerCall, size, aspectRatio, seed, providerOptions, maxRetries: maxRetriesArg, abortSignal, headers, }: {
+declare function generateImage({ model: modelArg, prompt, n, maxImagesPerCall, size, aspectRatio, seed, providerOptions, maxRetries: maxRetriesArg, abortSignal, headers, }: {
     /**
   The image model to use.
        */
-    model: ImageModelV3;
+    model: ImageModel;
     /**
   The prompt that should be used to generate the image.
      */

package/dist/index.d.ts

@@ -3,7 +3,7 @@ import * as _ai_sdk_provider_utils from '@ai-sdk/provider-utils';
 import { Tool, InferToolInput, InferToolOutput, FlexibleSchema, InferSchema, AssistantModelMessage, ToolModelMessage, ReasoningPart, ModelMessage, SystemModelMessage, UserModelMessage, ProviderOptions, IdGenerator, ToolCall, MaybePromiseLike, TextPart, FilePart, Resolvable, FetchFunction, DataContent } from '@ai-sdk/provider-utils';
 export { AssistantContent, AssistantModelMessage, DataContent, FilePart, FlexibleSchema, IdGenerator, ImagePart, InferSchema, InferToolInput, InferToolOutput, ModelMessage, Schema, SystemModelMessage, TextPart, Tool, ToolApprovalRequest, ToolApprovalResponse, ToolCallOptions, ToolCallPart, ToolContent, ToolExecuteFunction, ToolModelMessage, ToolResultPart, UserContent, UserModelMessage, asSchema, createIdGenerator, dynamicTool, generateId, jsonSchema, parseJsonEventStream, tool, zodSchema } from '@ai-sdk/provider-utils';
 import * as _ai_sdk_provider from '@ai-sdk/provider';
-import { EmbeddingModelV3, EmbeddingModelV2, EmbeddingModelV3Embedding, ImageModelV3, ImageModelV3CallWarning, ImageModelV3ProviderMetadata, JSONValue as JSONValue$1, LanguageModelV3, LanguageModelV2, LanguageModelV3FinishReason, LanguageModelV3CallWarning, LanguageModelV3Source, LanguageModelV3Middleware, EmbeddingModelV3Middleware, RerankingModelV3, SharedV3ProviderMetadata, SpeechModelV3, SpeechModelV2, SpeechModelV3CallWarning, TranscriptionModelV3, TranscriptionModelV2, TranscriptionModelV3CallWarning, LanguageModelV3Usage, ImageModelV3Usage, LanguageModelV3CallOptions, AISDKError, LanguageModelV3ToolCall, JSONSchema7, JSONParseError, TypeValidationError, JSONObject, SharedV3Warning, EmbeddingModelCallOptions, ProviderV3, ProviderV2, NoSuchModelError } from '@ai-sdk/provider';
+import { EmbeddingModelV3, EmbeddingModelV2, EmbeddingModelV3Embedding, ImageModelV3, ImageModelV2, ImageModelV3CallWarning, ImageModelV2CallWarning, ImageModelV3ProviderMetadata, ImageModelV2ProviderMetadata, JSONValue as JSONValue$1, LanguageModelV3, LanguageModelV2, LanguageModelV3FinishReason, LanguageModelV3CallWarning, LanguageModelV3Source, LanguageModelV3Middleware, EmbeddingModelV3Middleware, RerankingModelV3, SharedV3ProviderMetadata, SpeechModelV3, SpeechModelV2, SpeechModelV3CallWarning, TranscriptionModelV3, TranscriptionModelV2, TranscriptionModelV3CallWarning, LanguageModelV3Usage, ImageModelV3Usage, LanguageModelV3CallOptions, AISDKError, LanguageModelV3ToolCall, JSONSchema7, JSONParseError, TypeValidationError, JSONObject, SharedV3Warning, EmbeddingModelCallOptions, ProviderV3, ProviderV2, NoSuchModelError } from '@ai-sdk/provider';
 export { AISDKError, APICallError, EmptyResponseBodyError, InvalidPromptError, InvalidResponseDataError, JSONParseError, JSONSchema7, LoadAPIKeyError, LoadSettingError, NoContentGeneratedError, NoSuchModelError, TooManyEmbeddingValuesForCallError, TypeValidationError, UnsupportedFunctionalityError } from '@ai-sdk/provider';
 import { ServerResponse } from 'node:http';
 import { AttributeValue, Tracer } from '@opentelemetry/api';
@@ -22,16 +22,16 @@ type Embedding = EmbeddingModelV3Embedding;
 /**
 Image model that is used by the AI SDK Core functions.
   */
-type ImageModel = ImageModelV3;
+type ImageModel = string | ImageModelV3 | ImageModelV2;
 /**
 Warning from the model provider for this call. The call will proceed, but e.g.
 some settings might not be supported, which can lead to suboptimal results.
   */
-type ImageGenerationWarning = ImageModelV3CallWarning;
+type ImageGenerationWarning = ImageModelV3CallWarning | ImageModelV2CallWarning;
 /**
 Metadata from the model provider for this call
   */
-type ImageModelProviderMetadata = ImageModelV3ProviderMetadata;
+type ImageModelProviderMetadata = ImageModelV3ProviderMetadata | ImageModelV2ProviderMetadata;
 
 type ImageModelResponseMetadata = {
     /**
@@ -3951,11 +3951,11 @@ as body parameters.
 
 @returns A result object that contains the generated images.
  */
-declare function generateImage({ model, prompt, n, maxImagesPerCall, size, aspectRatio, seed, providerOptions, maxRetries: maxRetriesArg, abortSignal, headers, }: {
+declare function generateImage({ model: modelArg, prompt, n, maxImagesPerCall, size, aspectRatio, seed, providerOptions, maxRetries: maxRetriesArg, abortSignal, headers, }: {
     /**
   The image model to use.
        */
-    model: ImageModelV3;
+    model: ImageModel;
     /**
   The prompt that should be used to generate the image.
      */

package/dist/index.js

@@ -558,6 +558,20 @@ function asEmbeddingModelV3(model) {
   });
 }
 
+// src/model/as-image-model-v3.ts
+function asImageModelV3(model) {
+  if (model.specificationVersion === "v3") {
+    return model;
+  }
+  return new Proxy(model, {
+    get(target, prop) {
+      if (prop === "specificationVersion")
+        return "v3";
+      return target[prop];
+    }
+  });
+}
+
 // src/model/as-language-model-v3.ts
 function asLanguageModelV3(model) {
   if (model.specificationVersion === "v3") {
@@ -661,6 +675,20 @@ function resolveSpeechModel(model) {
   }
   return (_b = (_a15 = getGlobalProvider()).speechModel) == null ? void 0 : _b.call(_a15, model);
 }
+function resolveImageModel(model) {
+  if (typeof model !== "string") {
+    if (model.specificationVersion !== "v3" && model.specificationVersion !== "v2") {
+      const unsupportedModel = model;
+      throw new UnsupportedModelVersionError({
+        version: unsupportedModel.specificationVersion,
+        provider: unsupportedModel.provider,
+        modelId: unsupportedModel.modelId
+      });
+    }
+    return asImageModelV3(model);
+  }
+  return getGlobalProvider().imageModel(model);
+}
 function getGlobalProvider() {
   var _a15;
   return (_a15 = globalThis.AI_SDK_DEFAULT_PROVIDER) != null ? _a15 : import_gateway.gateway;
@@ -858,7 +886,7 @@ function detectMediaType({
 var import_provider_utils2 = require("@ai-sdk/provider-utils");
 
 // src/version.ts
-var VERSION = true ? "6.0.0-beta.103" : "0.0.0-test";
+var VERSION = true ? "6.0.0-beta.104" : "0.0.0-test";
 
 // src/util/download/download.ts
 var download = async ({ url }) => {
@@ -7893,7 +7921,7 @@ var DefaultEmbedManyResult = class {
 // src/generate-image/generate-image.ts
 var import_provider_utils23 = require("@ai-sdk/provider-utils");
 async function generateImage({
-  model,
+  model: modelArg,
   prompt,
   n = 1,
   maxImagesPerCall,
@@ -7906,13 +7934,7 @@ async function generateImage({
   headers
 }) {
   var _a15, _b;
-  if (model.specificationVersion !== "v3") {
-    throw new UnsupportedModelVersionError({
-      version: model.specificationVersion,
-      provider: model.provider,
-      modelId: model.modelId
-    });
-  }
+  const model = resolveImageModel(modelArg);
   const headersWithUserAgent = (0, import_provider_utils23.withUserAgentSuffix)(
     headers != null ? headers : {},
     `ai/${VERSION}`
@@ -7976,10 +7998,26 @@ async function generateImage({
     }
     if (result.providerMetadata) {
       for (const [providerName, metadata] of Object.entries(result.providerMetadata)) {
-        (_b = providerMetadata[providerName]) != null ? _b : providerMetadata[providerName] = { images: [] };
-        providerMetadata[providerName].images.push(
-          ...result.providerMetadata[providerName].images
-        );
+        if (providerName === "gateway") {
+          const currentEntry = providerMetadata[providerName];
+          if (currentEntry != null && typeof currentEntry === "object") {
+            providerMetadata[providerName] = {
+              ...currentEntry,
+              ...metadata
+            };
+          } else {
+            providerMetadata[providerName] = metadata;
+          }
+          const imagesValue = providerMetadata[providerName].images;
+          if (Array.isArray(imagesValue) && imagesValue.length === 0) {
+            delete providerMetadata[providerName].images;
+          }
+        } else {
+          (_b = providerMetadata[providerName]) != null ? _b : providerMetadata[providerName] = { images: [] };
+          providerMetadata[providerName].images.push(
+            ...result.providerMetadata[providerName].images
+          );
+        }
       }
     }
     responses.push(result.response);
@@ -10015,20 +10053,6 @@ var doWrap2 = ({
   };
 };
 
-// src/model/as-image-model-v3.ts
-function asImageModelV3(model) {
-  if (model.specificationVersion === "v3") {
-    return model;
-  }
-  return new Proxy(model, {
-    get(target, prop) {
-      if (prop === "specificationVersion")
-        return "v3";
-      return target[prop];
-    }
-  });
-}
-
 // src/model/as-provider-v3.ts
 function asProviderV3(provider) {
   if ("specificationVersion" in provider && provider.specificationVersion === "v3") {