@ai-sdk/google - npm Package Compare versions

Comparing version 3.0.47 to 3.0.48

CHANGELOG.md

 # @ai-sdk/google
 
+## 3.0.48
+
+### Patch Changes
+
+- 35c46d1: feat(provider/google): support multimodal content parts in embedding provider options
+
 ## 3.0.47
@@ -4,0 +10,0 @@

dist/index.d.mts

@@ -189,2 +189,10 @@ import * as _ai_sdk_provider_utils from '@ai-sdk/provider-utils';
     taskType?: "SEMANTIC_SIMILARITY" | "CLASSIFICATION" | "CLUSTERING" | "RETRIEVAL_DOCUMENT" | "RETRIEVAL_QUERY" | "QUESTION_ANSWERING" | "FACT_VERIFICATION" | "CODE_RETRIEVAL_QUERY" | undefined;
+    content?: ({
+        text: string;
+    } | {
+        inlineData: {
+            mimeType: string;
+            data: string;
+        };
+    })[] | undefined;
 }>;
@@ -191,0 +199,0 @@ type GoogleEmbeddingModelOptions = InferSchema<typeof googleEmbeddingModelOptions>;

dist/index.d.ts

@@ -189,2 +189,10 @@ import * as _ai_sdk_provider_utils from '@ai-sdk/provider-utils';
     taskType?: "SEMANTIC_SIMILARITY" | "CLASSIFICATION" | "CLUSTERING" | "RETRIEVAL_DOCUMENT" | "RETRIEVAL_QUERY" | "QUESTION_ANSWERING" | "FACT_VERIFICATION" | "CODE_RETRIEVAL_QUERY" | undefined;
+    content?: ({
+        text: string;
+    } | {
+        inlineData: {
+            mimeType: string;
+            data: string;
+        };
+    })[] | undefined;
 }>;
@@ -191,0 +199,0 @@ type GoogleEmbeddingModelOptions = InferSchema<typeof googleEmbeddingModelOptions>;

package.json

 {
   "name": "@ai-sdk/google",
-  "version": "3.0.47",
+  "version": "3.0.48",
   "license": "Apache-2.0",
@@ -47,4 +47,4 @@ "sideEffects": false,
     "zod": "3.25.76",
-    "@ai-sdk/test-server": "1.0.3",
-    "@vercel/ai-tsconfig": "0.0.0"
+    "@vercel/ai-tsconfig": "0.0.0",
+    "@ai-sdk/test-server": "1.0.3"
   },
@@ -51,0 +51,0 @@ "peerDependencies": {

src/google-generative-ai-embedding-model.ts

@@ -77,4 +77,12 @@ import {
 
-    // For single embeddings, use the single endpoint (ratelimits, etc.)
+    const multimodalContent = googleOptions?.content ?? [];
+
+    // For single embeddings, use the single endpoint
     if (values.length === 1) {
+      const textPart = values[0] ? [{ text: values[0] }] : [];
+      const parts =
+        multimodalContent.length > 0
+          ? [...textPart, ...multimodalContent]
+          : [{ text: values[0] }];
+
       const {
@@ -90,3 +98,3 @@ responseHeaders,
           content: {
-            parts: [{ text: values[0] }],
+            parts,
           },
@@ -112,2 +120,4 @@ outputDimensionality: googleOptions?.outputDimensionality,
 
+    // For multiple values, use the batch endpoint
+    // If multimodal content is provided, merge it into each request's parts
     const {
@@ -121,8 +131,17 @@ responseHeaders,
       body: {
-        requests: values.map(value => ({
-          model: `models/${this.modelId}`,
-          content: { role: 'user', parts: [{ text: value }] },
-          outputDimensionality: googleOptions?.outputDimensionality,
-          taskType: googleOptions?.taskType,
-        })),
+        requests: values.map(value => {
+          const textPart = value ? [{ text: value }] : [];
+          return {
+            model: `models/${this.modelId}`,
+            content: {
+              role: 'user',
+              parts:
+                multimodalContent.length > 0
+                  ? [...textPart, ...multimodalContent]
+                  : [{ text: value }],
+            },
+            outputDimensionality: googleOptions?.outputDimensionality,
+            taskType: googleOptions?.taskType,
+          };
+        }),
       },
@@ -129,0 +148,0 @@ failedResponseHandler: googleFailedResponseHandler,

src/google-generative-ai-embedding-options.ts

@@ -12,2 +12,12 @@ import {
 
+const googleEmbeddingContentPartSchema = z.union([
+  z.object({ text: z.string() }),
+  z.object({
+    inlineData: z.object({
+      mimeType: z.string(),
+      data: z.string(),
+    }),
+  }),
+]);
+
 export const googleEmbeddingModelOptions = lazySchema(() =>
@@ -46,2 +56,9 @@ zodSchema(
         .optional(),
+
+      /**
+       * Optional. Multimodal content parts for embedding non-text content
+       * (images, video, PDF, audio). When provided, these parts are used
+       * instead of the text values in the embedding request.
+       */
+      content: z.array(googleEmbeddingContentPartSchema).min(1).optional(),
     }),
@@ -48,0 +65,0 @@ ),

New alerts

Network access

Supply chain risk

This module accesses the network.

Found 1 instance in 1 package

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Found 1 instance in 1 package

URL strings

Supply chain risk

Package contains fragments of external URLs or IP addresses, which the package may be accessing at runtime.

Found 1 instance in 1 package

Fixed alerts

Network access

Supply chain risk

This module accesses the network.

Found 1 instance in 1 package

Long strings

Supply chain risk

Contains long string literals, which may be a sign of obfuscated or packed code.

Found 1 instance in 1 package

URL strings

Supply chain risk

Package contains fragments of external URLs or IP addresses, which the package may be accessing at runtime.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

1105914

0.65%

Lines of code

12250

0.74%

No dependency changes