@ai-sdk/alibaba 2.0.0-beta.6 → 2.0.0-beta.61

package/src/alibaba-provider.ts

@@ -1,37 +1,26 @@
 import {
+  NoSuchModelError,
+  type EmbeddingModelV4,
   type Experimental_VideoModelV4,
   type LanguageModelV4,
-  NoSuchModelError,
   type ProviderV4,
 } from '@ai-sdk/provider';
 import {
-  createJsonErrorResponseHandler,
-  type FetchFunction,
   loadApiKey,
   withoutTrailingSlash,
   withUserAgentSuffix,
+  type FetchFunction,
 } from '@ai-sdk/provider-utils';
-import { z } from 'zod/v4';
-import { AlibabaLanguageModel } from './alibaba-chat-language-model';
-import type { AlibabaChatModelId } from './alibaba-chat-options';
+import { AlibabaChatLanguageModel } from './alibaba-chat-language-model';
+import type { AlibabaChatModelId } from './alibaba-chat-language-model-options';
+import { AlibabaEmbeddingModel } from './alibaba-embedding-model';
+import type { AlibabaEmbeddingModelId } from './alibaba-embedding-model-options';
 import { AlibabaVideoModel } from './alibaba-video-model';
 import type { AlibabaVideoModelId } from './alibaba-video-settings';
 import { VERSION } from './version';
 
-export type AlibabaErrorData = z.infer<typeof alibabaErrorDataSchema>;
-
-const alibabaErrorDataSchema = z.object({
-  error: z.object({
-    message: z.string(),
-    code: z.string().nullish(),
-    type: z.string().nullish(),
-  }),
-});
-
-export const alibabaFailedResponseHandler = createJsonErrorResponseHandler({
-  errorSchema: alibabaErrorDataSchema,
-  errorToMessage: data => data.error.message,
-});
+export type { AlibabaErrorData } from './alibaba-error';
+export { alibabaFailedResponseHandler } from './alibaba-error';
 
 export interface AlibabaProvider extends ProviderV4 {
   (modelId: AlibabaChatModelId): LanguageModelV4;
@@ -46,6 +35,16 @@ export interface AlibabaProvider extends ProviderV4 {
    */
   chatModel(modelId: AlibabaChatModelId): LanguageModelV4;
 
+  /**
+   * Creates a model for text embeddings.
+   */
+  embedding(modelId: AlibabaEmbeddingModelId): EmbeddingModelV4;
+
+  /**
+   * Creates a model for text embeddings.
+   */
+  embeddingModel(modelId: AlibabaEmbeddingModelId): EmbeddingModelV4;
+
   /**
    * Creates a model for video generation.
    */
@@ -71,6 +70,13 @@ export interface AlibabaProviderSettings {
    */
   videoBaseURL?: string;
 
+  /**
+   * Use a different URL prefix for embedding API calls.
+   * The embedding API uses the DashScope native endpoint (not the OpenAI-compatible endpoint).
+   * The default prefix is `https://dashscope-intl.aliyuncs.com/api/v1`.
+   */
+  embeddingBaseURL?: string;
+
   /**
    * API key that is being sent using the `Authorization` header.
    * It defaults to the `ALIBABA_API_KEY` environment variable.
@@ -111,6 +117,10 @@ export function createAlibaba(
     withoutTrailingSlash(options.videoBaseURL) ??
     'https://dashscope-intl.aliyuncs.com';
 
+  const embeddingBaseURL =
+    withoutTrailingSlash(options.embeddingBaseURL) ??
+    'https://dashscope-intl.aliyuncs.com/api/v1';
+
   const getHeaders = () =>
     withUserAgentSuffix(
       {
@@ -125,7 +135,7 @@ export function createAlibaba(
     );
 
   const createLanguageModel = (modelId: AlibabaChatModelId) =>
-    new AlibabaLanguageModel(modelId, {
+    new AlibabaChatLanguageModel(modelId, {
       provider: 'alibaba.chat',
       baseURL,
       headers: getHeaders,
@@ -133,6 +143,14 @@ export function createAlibaba(
       includeUsage: options.includeUsage ?? true,
     });
 
+  const createEmbeddingModel = (modelId: AlibabaEmbeddingModelId) =>
+    new AlibabaEmbeddingModel(modelId, {
+      provider: 'alibaba.embedding',
+      baseURL: embeddingBaseURL,
+      headers: getHeaders,
+      fetch: options.fetch,
+    });
+
   const createVideoModel = (modelId: AlibabaVideoModelId) =>
     new AlibabaVideoModel(modelId, {
       provider: 'alibaba.video',
@@ -154,6 +172,8 @@ export function createAlibaba(
   provider.specificationVersion = 'v4' as const;
   provider.languageModel = createLanguageModel;
   provider.chatModel = createLanguageModel;
+  provider.embedding = createEmbeddingModel;
+  provider.embeddingModel = createEmbeddingModel;
   provider.video = createVideoModel;
   provider.videoModel = createVideoModel;
 
@@ -161,10 +181,6 @@ export function createAlibaba(
     throw new NoSuchModelError({ modelId, modelType: 'imageModel' });
   };
 
-  provider.embeddingModel = (modelId: string) => {
-    throw new NoSuchModelError({ modelId, modelType: 'embeddingModel' });
-  };
-
   return provider;
 }
 

package/src/alibaba-video-model-options.ts

@@ -0,0 +1,46 @@
+import { lazySchema, zodSchema } from '@ai-sdk/provider-utils';
+import { z } from 'zod/v4';
+
+export type AlibabaVideoModelOptions = {
+  /** Negative prompt to specify what to avoid (max 500 chars). */
+  negativePrompt?: string | null;
+  /** URL to audio file for audio-video sync (WAV/MP3, 3-30s, max 15MB). */
+  audioUrl?: string | null;
+  /** Enable prompt extension/rewriting for better generation. Defaults to true. */
+  promptExtend?: boolean | null;
+  /** Shot type: 'single' for single-shot or 'multi' for multi-shot narrative. */
+  shotType?: 'single' | 'multi' | null;
+  /** Whether to add watermark to generated video. Defaults to false. */
+  watermark?: boolean | null;
+  /** Enable audio generation (for I2V/R2V models). */
+  audio?: boolean | null;
+  /**
+   * Reference URLs for reference-to-video mode.
+   * Array of URLs to images (0-5) and/or videos (0-3), max 5 total.
+   * Use character identifiers (character1, character2) in prompts to reference them.
+   */
+  referenceUrls?: string[] | null;
+  /** Polling interval in milliseconds. Defaults to 5000 (5 seconds). */
+  pollIntervalMs?: number | null;
+  /** Maximum wait time in milliseconds for video generation. Defaults to 600000 (10 minutes). */
+  pollTimeoutMs?: number | null;
+  [key: string]: unknown;
+};
+
+export const alibabaVideoModelOptionsSchema = lazySchema(() =>
+  zodSchema(
+    z
+      .object({
+        negativePrompt: z.string().nullish(),
+        audioUrl: z.string().nullish(),
+        promptExtend: z.boolean().nullish(),
+        shotType: z.enum(['single', 'multi']).nullish(),
+        watermark: z.boolean().nullish(),
+        audio: z.boolean().nullish(),
+        referenceUrls: z.array(z.string()).nullish(),
+        pollIntervalMs: z.number().positive().nullish(),
+        pollTimeoutMs: z.number().positive().nullish(),
+      })
+      .passthrough(),
+  ),
+);

package/src/alibaba-video-model.ts

@@ -9,62 +9,20 @@ import {
   createJsonErrorResponseHandler,
   createJsonResponseHandler,
   delay,
-  type FetchFunction,
   getFromApi,
-  lazySchema,
   parseProviderOptions,
   postJsonToApi,
-  type Resolvable,
   resolve,
-  zodSchema,
+  type FetchFunction,
+  type Resolvable,
 } from '@ai-sdk/provider-utils';
 import { z } from 'zod/v4';
+import {
+  alibabaVideoModelOptionsSchema,
+  type AlibabaVideoModelOptions,
+} from './alibaba-video-model-options';
 import type { AlibabaVideoModelId } from './alibaba-video-settings';
 
-export type AlibabaVideoModelOptions = {
-  /** Negative prompt to specify what to avoid (max 500 chars). */
-  negativePrompt?: string | null;
-  /** URL to audio file for audio-video sync (WAV/MP3, 3-30s, max 15MB). */
-  audioUrl?: string | null;
-  /** Enable prompt extension/rewriting for better generation. Defaults to true. */
-  promptExtend?: boolean | null;
-  /** Shot type: 'single' for single-shot or 'multi' for multi-shot narrative. */
-  shotType?: 'single' | 'multi' | null;
-  /** Whether to add watermark to generated video. Defaults to false. */
-  watermark?: boolean | null;
-  /** Enable audio generation (for I2V/R2V models). */
-  audio?: boolean | null;
-  /**
-   * Reference URLs for reference-to-video mode.
-   * Array of URLs to images (0-5) and/or videos (0-3), max 5 total.
-   * Use character identifiers (character1, character2) in prompts to reference them.
-   */
-  referenceUrls?: string[] | null;
-  /** Polling interval in milliseconds. Defaults to 5000 (5 seconds). */
-  pollIntervalMs?: number | null;
-  /** Maximum wait time in milliseconds for video generation. Defaults to 600000 (10 minutes). */
-  pollTimeoutMs?: number | null;
-  [key: string]: unknown;
-};
-
-const alibabaVideoModelOptionsSchema = lazySchema(() =>
-  zodSchema(
-    z
-      .object({
-        negativePrompt: z.string().nullish(),
-        audioUrl: z.string().nullish(),
-        promptExtend: z.boolean().nullish(),
-        shotType: z.enum(['single', 'multi']).nullish(),
-        watermark: z.boolean().nullish(),
-        audio: z.boolean().nullish(),
-        referenceUrls: z.array(z.string()).nullish(),
-        pollIntervalMs: z.number().positive().nullish(),
-        pollTimeoutMs: z.number().positive().nullish(),
-      })
-      .passthrough(),
-  ),
-);
-
 interface AlibabaVideoModelConfig {
   provider: string;
   baseURL: string;
@@ -241,8 +199,9 @@ export class AlibabaVideoModel implements Experimental_VideoModelV4 {
     if (alibabaOptions?.watermark != null) {
       parameters.watermark = alibabaOptions.watermark;
     }
-    if (alibabaOptions?.audio != null) {
-      parameters.audio = alibabaOptions.audio;
+    const audio = options.generateAudio ?? alibabaOptions?.audio;
+    if (audio != null) {
+      parameters.audio = audio;
     }
 
     // Warn about unsupported standard options

package/src/convert-to-alibaba-chat-messages.ts

@@ -1,22 +1,28 @@
 import {
-  type LanguageModelV4DataContent,
-  type LanguageModelV4Prompt,
   UnsupportedFunctionalityError,
+  type LanguageModelV4FilePart,
+  type LanguageModelV4Prompt,
 } from '@ai-sdk/provider';
-import { convertToBase64 } from '@ai-sdk/provider-utils';
+import {
+  convertToBase64,
+  getTopLevelMediaType,
+  resolveFullMediaType,
+} from '@ai-sdk/provider-utils';
 import type { AlibabaChatPrompt } from './alibaba-chat-prompt';
 import type { CacheControlValidator } from './get-cache-control';
 
-function formatImageUrl({
-  data,
-  mediaType,
-}: {
-  data: LanguageModelV4DataContent;
-  mediaType: string;
-}): string {
-  return data instanceof URL
-    ? data.toString()
-    : `data:${mediaType};base64,${convertToBase64(data as Uint8Array)}`;
+function formatImageUrl({ part }: { part: LanguageModelV4FilePart }): string {
+  if (part.data.type === 'url') {
+    return part.data.url.toString();
+  }
+
+  if (part.data.type === 'data') {
+    return `data:${resolveFullMediaType({ part })};base64,${convertToBase64(part.data.data)}`;
+  }
+
+  throw new UnsupportedFunctionalityError({
+    functionality: `file part data type ${part.data.type}`,
+  });
 }
 
 export function convertToAlibabaChatMessages({
@@ -53,26 +59,13 @@ export function convertToAlibabaChatMessages({
       }
 
       case 'user': {
-        const isSinglePart = content.length === 1;
-
-        if (
-          isSinglePart &&
-          content[0].type === 'text' &&
-          !messageCacheControl
-        ) {
-          messages.push({
-            role: 'user',
-            content: content[0].text,
-          });
-          break;
-        }
-
         messages.push({
           role: 'user',
-          content: content.map(part => {
-            const partCacheControl = isSinglePart
-              ? messageCacheControl
-              : cacheControlValidator?.getCacheControl(part.providerOptions);
+          content: content.map((part, index) => {
+            const isLastPart = index === content.length - 1;
+            const partCacheControl =
+              cacheControlValidator?.getCacheControl(part.providerOptions) ??
+              (isLastPart ? messageCacheControl : undefined);
 
             switch (part.type) {
               case 'text': {
@@ -86,25 +79,35 @@ export function convertToAlibabaChatMessages({
               }
 
               case 'file': {
-                if (part.mediaType.startsWith('image/')) {
-                  const mediaType =
-                    part.mediaType === 'image/*'
-                      ? 'image/jpeg'
-                      : part.mediaType;
-
-                  return {
-                    type: 'image_url',
-                    image_url: {
-                      url: formatImageUrl({ data: part.data, mediaType }),
-                    },
-                    ...(partCacheControl
-                      ? { cache_control: partCacheControl }
-                      : {}),
-                  };
-                } else {
-                  throw new UnsupportedFunctionalityError({
-                    functionality: 'Only image file parts are supported',
-                  });
+                switch (part.data.type) {
+                  case 'reference': {
+                    throw new UnsupportedFunctionalityError({
+                      functionality: 'file parts with provider references',
+                    });
+                  }
+                  case 'text': {
+                    throw new UnsupportedFunctionalityError({
+                      functionality: 'text file parts',
+                    });
+                  }
+                  case 'url':
+                  case 'data': {
+                    if (getTopLevelMediaType(part.mediaType) === 'image') {
+                      return {
+                        type: 'image_url',
+                        image_url: {
+                          url: formatImageUrl({ part }),
+                        },
+                        ...(partCacheControl
+                          ? { cache_control: partCacheControl }
+                          : {}),
+                      };
+                    } else {
+                      throw new UnsupportedFunctionalityError({
+                        functionality: 'Only image file parts are supported',
+                      });
+                    }
+                  }
                 }
               }
             }
@@ -163,17 +166,15 @@ export function convertToAlibabaChatMessages({
           r => r.type !== 'tool-approval-response',
         );
 
-        const isSinglePart = toolResponses.length === 1;
-
         for (let i = 0; i < toolResponses.length; i++) {
           const toolResponse = toolResponses[i];
           const output = toolResponse.output;
+          const isLastPart = i === toolResponses.length - 1;
 
-          const partCacheControl = isSinglePart
-            ? messageCacheControl
-            : cacheControlValidator?.getCacheControl(
-                (toolResponse as any).providerOptions,
-              );
+          const partCacheControl =
+            cacheControlValidator?.getCacheControl(
+              toolResponse.providerOptions,
+            ) ?? (isLastPart ? messageCacheControl : undefined);
 
           let contentValue: string;
           switch (output.type) {
@@ -182,7 +183,7 @@ export function convertToAlibabaChatMessages({
               contentValue = output.value;
               break;
             case 'execution-denied':
-              contentValue = output.reason ?? 'Tool execution denied.';
+              contentValue = output.reason ?? 'Tool call execution denied.';
               break;
             case 'content':
             case 'json':

package/src/index.ts

@@ -1,10 +1,16 @@
 export type {
   AlibabaChatModelId,
-  AlibabaLanguageModelOptions,
-  /** @deprecated Use `AlibabaLanguageModelOptions` instead. */
-  AlibabaLanguageModelOptions as AlibabaProviderOptions,
-} from './alibaba-chat-options';
+  AlibabaLanguageModelChatOptions,
+  /** @deprecated Use `AlibabaLanguageModelChatOptions` instead. */
+  AlibabaLanguageModelChatOptions as AlibabaLanguageModelOptions,
+  /** @deprecated Use `AlibabaLanguageModelChatOptions` instead. */
+  AlibabaLanguageModelChatOptions as AlibabaProviderOptions,
+} from './alibaba-chat-language-model-options';
 export type { AlibabaCacheControl } from './alibaba-chat-prompt';
+export type {
+  AlibabaEmbeddingModelId,
+  AlibabaEmbeddingModelOptions,
+} from './alibaba-embedding-model-options';
 export {
   type AlibabaProvider,
   type AlibabaProviderSettings,
@@ -15,7 +21,7 @@ export type {
   AlibabaVideoModelOptions,
   /** @deprecated Use `AlibabaVideoModelOptions` instead. */
   AlibabaVideoModelOptions as AlibabaVideoProviderOptions,
-} from './alibaba-video-model';
+} from './alibaba-video-model-options';
 export type { AlibabaVideoModelId } from './alibaba-video-settings';
 export type { AlibabaUsage } from './convert-alibaba-usage';
 export { VERSION } from './version';

package/src/version.ts

@@ -1,3 +1,6 @@
-declare const __PACKAGE_VERSION__: string;
-
-export const VERSION = __PACKAGE_VERSION__;
+// Version string of this package injected at build time.
+declare const __PACKAGE_VERSION__: string | undefined;
+export const VERSION: string =
+  typeof __PACKAGE_VERSION__ !== 'undefined'
+    ? __PACKAGE_VERSION__
+    : '0.0.0-test';

package/dist/index.d.mts

@@ -1,119 +0,0 @@
-import { z } from 'zod/v4';
-import { FetchFunction } from '@ai-sdk/provider-utils';
-import { ProviderV4, LanguageModelV4, Experimental_VideoModelV4 } from '@ai-sdk/provider';
-
-type AlibabaChatModelId = 'qwen3-max' | 'qwen3-max-preview' | 'qwen-plus' | 'qwen-plus-latest' | 'qwen-flash' | 'qwen-turbo' | 'qwen-turbo-latest' | 'qwen3-235b-a22b' | 'qwen3-32b' | 'qwen3-30b-a3b' | 'qwen3-14b' | 'qwen3-next-80b-a3b-thinking' | 'qwen3-235b-a22b-thinking-2507' | 'qwen3-30b-a3b-thinking-2507' | 'qwq-plus' | 'qwq-plus-latest' | 'qwq-32b' | 'qwen-coder' | 'qwen3-coder-plus' | 'qwen3-coder-flash' | (string & {});
-declare const alibabaLanguageModelOptions: z.ZodObject<{
-    enableThinking: z.ZodOptional<z.ZodBoolean>;
-    thinkingBudget: z.ZodOptional<z.ZodNumber>;
-    parallelToolCalls: z.ZodOptional<z.ZodBoolean>;
-}, z.core.$strip>;
-type AlibabaLanguageModelOptions = z.infer<typeof alibabaLanguageModelOptions>;
-
-type AlibabaCacheControl = {
-    type: string;
-};
-
-type AlibabaVideoModelId = 'wan2.6-t2v' | 'wan2.5-t2v-preview' | 'wan2.6-i2v' | 'wan2.6-i2v-flash' | 'wan2.6-r2v' | 'wan2.6-r2v-flash' | (string & {});
-
-interface AlibabaProvider extends ProviderV4 {
-    (modelId: AlibabaChatModelId): LanguageModelV4;
-    /**
-     * Creates a model for text generation.
-     */
-    languageModel(modelId: AlibabaChatModelId): LanguageModelV4;
-    /**
-     * Creates a chat model for text generation.
-     */
-    chatModel(modelId: AlibabaChatModelId): LanguageModelV4;
-    /**
-     * Creates a model for video generation.
-     */
-    video(modelId: AlibabaVideoModelId): Experimental_VideoModelV4;
-    /**
-     * Creates a model for video generation.
-     */
-    videoModel(modelId: AlibabaVideoModelId): Experimental_VideoModelV4;
-}
-interface AlibabaProviderSettings {
-    /**
-     * Use a different URL prefix for API calls, e.g. to use proxy servers or regional endpoints.
-     * The default prefix is `https://dashscope-intl.aliyuncs.com/compatible-mode/v1`.
-     */
-    baseURL?: string;
-    /**
-     * Use a different URL prefix for video generation API calls.
-     * The video API uses the DashScope native endpoint (not the OpenAI-compatible endpoint).
-     * The default prefix is `https://dashscope-intl.aliyuncs.com`.
-     */
-    videoBaseURL?: string;
-    /**
-     * API key that is being sent using the `Authorization` header.
-     * It defaults to the `ALIBABA_API_KEY` environment variable.
-     */
-    apiKey?: string;
-    /**
-     * Custom headers to include in the requests.
-     */
-    headers?: Record<string, string>;
-    /**
-     * Custom fetch implementation. You can use it as a middleware to intercept requests,
-     * or to provide a custom fetch implementation for e.g. testing.
-     */
-    fetch?: FetchFunction;
-    /**
-     * Include usage information in streaming responses.
-     * When enabled, token usage will be included in the final chunk.
-     *
-     * @default true
-     */
-    includeUsage?: boolean;
-}
-/**
- * Create an Alibaba Cloud (Qwen) provider instance.
- */
-declare function createAlibaba(options?: AlibabaProviderSettings): AlibabaProvider;
-declare const alibaba: AlibabaProvider;
-
-type AlibabaVideoModelOptions = {
-    /** Negative prompt to specify what to avoid (max 500 chars). */
-    negativePrompt?: string | null;
-    /** URL to audio file for audio-video sync (WAV/MP3, 3-30s, max 15MB). */
-    audioUrl?: string | null;
-    /** Enable prompt extension/rewriting for better generation. Defaults to true. */
-    promptExtend?: boolean | null;
-    /** Shot type: 'single' for single-shot or 'multi' for multi-shot narrative. */
-    shotType?: 'single' | 'multi' | null;
-    /** Whether to add watermark to generated video. Defaults to false. */
-    watermark?: boolean | null;
-    /** Enable audio generation (for I2V/R2V models). */
-    audio?: boolean | null;
-    /**
-     * Reference URLs for reference-to-video mode.
-     * Array of URLs to images (0-5) and/or videos (0-3), max 5 total.
-     * Use character identifiers (character1, character2) in prompts to reference them.
-     */
-    referenceUrls?: string[] | null;
-    /** Polling interval in milliseconds. Defaults to 5000 (5 seconds). */
-    pollIntervalMs?: number | null;
-    /** Maximum wait time in milliseconds for video generation. Defaults to 600000 (10 minutes). */
-    pollTimeoutMs?: number | null;
-    [key: string]: unknown;
-};
-
-type AlibabaUsage = {
-    prompt_tokens?: number | null;
-    completion_tokens?: number | null;
-    total_tokens?: number | null;
-    prompt_tokens_details?: {
-        cached_tokens?: number | null;
-        cache_creation_input_tokens?: number | null;
-    } | null;
-    completion_tokens_details?: {
-        reasoning_tokens?: number | null;
-    } | null;
-};
-
-declare const VERSION: string;
-
-export { type AlibabaCacheControl, type AlibabaChatModelId, type AlibabaLanguageModelOptions, type AlibabaProvider, type AlibabaLanguageModelOptions as AlibabaProviderOptions, type AlibabaProviderSettings, type AlibabaUsage, type AlibabaVideoModelId, type AlibabaVideoModelOptions, type AlibabaVideoModelOptions as AlibabaVideoProviderOptions, VERSION, alibaba, createAlibaba };