feat (provider/{google, google-vertex}): expose type for validating provider options

.changeset/pink-deers-switch.md

@@ -0,0 +1,6 @@
+---
+'@ai-sdk/google-vertex': patch
+'@ai-sdk/google': patch
+---
+
+feat: add provider option schemas for vertex imagegen and google genai

content/providers/01-ai-sdk-providers/15-google-generative-ai.mdx

@@ -80,7 +80,7 @@ const model = google('gemini-1.5-pro-latest');
   e.g. `tunedModels/my-model`.
 </Note>
 
-Google Generative AI models support also some model specific settings that are not part of the [standard call settings](/docs/ai-sdk-core/settings).
+Google Generative AI also supports some model specific settings that are not part of the [standard call settings](/docs/ai-sdk-core/settings).
 You can pass them as an options argument:
 
 ```ts
@@ -132,6 +132,29 @@ The following optional settings are available for Google Generative AI models:
     - `BLOCK_ONLY_HIGH`
     - `BLOCK_NONE`
 
+Further configuration can be done using Google Generative AI provider options. You can validate the provider options using the `GoogleGenerativeAIProviderOptions` type.
+
+```ts
+import { google } from '@ai-sdk/google';
+import { GoogleGenerativeAIProviderOptions } from '@ai-sdk/google';
+import { generateText } from 'ai';
+
+const { text } = await generateText({
+  model: google('gemini-1.5-pro-latest'),
+  providerOptions: {
+    google: {
+      responseModalities: ['TEXT', 'IMAGE'],
+    } satisfies GoogleGenerativeAIProviderOptions,
+  },
+  // ...
+});
+```
+
+The following provider options are available:
+
+- **responseModalities** _string[]_
+  The modalities to use for the response. The following modalities are supported: `TEXT`, `IMAGE`. When not defined or empty, the model defaults to returning only text.
+
 You can use Google Generative AI language models to generate text with the `generateText` function:
 
 ```ts

content/providers/01-ai-sdk-providers/16-google-vertex.mdx

@@ -599,6 +599,41 @@ const { image } = await generateImage({
 });
 ```
 
+Further configuration can be done using Google Vertex provider options. You can validate the provider options using the `GoogleVertexImageProviderOptions` type.
+
+```ts
+import { vertex } from '@ai-sdk/google-vertex';
+import { GoogleVertexImageProviderOptions } from '@ai-sdk/google-vertex';
+import { generateImage } from 'ai';
+
+const { image } = await generateImage({
+  model: vertex.image('imagen-3.0-generate-001'),
+  providerOptions: {
+    vertex: {
+      negativePrompt: 'pixelated, blurry, low-quality',
+    } satisfies GoogleVertexImageProviderOptions,
+  },
+  // ...
+});
+```
+
+The following provider options are available:
+
+- **negativePrompt** _string_
+  A description of what to discourage in the generated images.
+
+- **personGeneration** `allow_adult` | `allow_all` | `dont_allow`
+  Whether to allow person generation. Defaults to `allow_adult`.
+
+- **safetySetting** `block_low_and_above` | `block_medium_and_above` | `block_only_high` | `block_none`
+  Whether to block unsafe content. Defaults to `block_medium_and_above`.
+
+- **addWatermark** _boolean_
+  Whether to add an invisible watermark to the generated images. Defaults to `true`.
+
+- **storageUri** _string_
+  Cloud Storage URI to store the generated images.
+
 <Note>
   Imagen models do not support the `size` parameter. Use the `aspectRatio`
   parameter instead.

examples/ai-core/src/generate-image/google-vertex.ts

@@ -1,7 +1,10 @@
-import { vertex } from '@ai-sdk/google-vertex';
+import {
+  GoogleVertexImageProviderOptions,
+  vertex,
+} from '@ai-sdk/google-vertex';
 import { experimental_generateImage as generateImage } from 'ai';
-import { presentImages } from '../lib/present-image';
 import 'dotenv/config';
+import { presentImages } from '../lib/present-image';
 
 async function main() {
   const { image } = await generateImage({
@@ -10,9 +13,8 @@ async function main() {
     aspectRatio: '1:1',
     providerOptions: {
       vertex: {
-        // https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/imagen-api#parameter_list
         addWatermark: false,
-      },
+      } satisfies GoogleVertexImageProviderOptions,
     },
   });
 

examples/ai-core/src/generate-text/google-image.ts

@@ -1,4 +1,4 @@
-import { google } from '@ai-sdk/google';
+import { google, GoogleGenerativeAIProviderOptions } from '@ai-sdk/google';
 import { generateText } from 'ai';
 import 'dotenv/config';
 import fs from 'node:fs';
@@ -15,6 +15,11 @@ async function main() {
         ],
       },
     ],
+    providerOptions: {
+      google: {
+        responseModalities: ['TEXT', 'IMAGE'],
+      } satisfies GoogleGenerativeAIProviderOptions,
+    },
   });
 
   console.log(result.text);

packages/google-vertex/src/google-vertex-image-model.test.ts

@@ -38,27 +38,6 @@ describe('GoogleVertexImageModel', () => {
       };
     }
 
-    it('should pass the correct parameters', async () => {
-      prepareJsonResponse();
-
-      await model.doGenerate({
-        prompt,
-        n: 2,
-        size: undefined,
-        aspectRatio: undefined,
-        seed: undefined,
-        providerOptions: { vertex: { aspectRatio: '1:1' } },
-      });
-
-      expect(await server.calls[0].requestBody).toStrictEqual({
-        instances: [{ prompt }],
-        parameters: {
-          sampleCount: 2,
-          aspectRatio: '1:1',
-        },
-      });
-    });
-
     it('should pass headers', async () => {
       prepareJsonResponse();
 
@@ -143,13 +122,9 @@ describe('GoogleVertexImageModel', () => {
         prompt: 'test prompt',
         n: 1,
         size: undefined,
-        aspectRatio: undefined,
+        aspectRatio: '16:9',
         seed: undefined,
-        providerOptions: {
-          vertex: {
-            aspectRatio: '16:9',
-          },
-        },
+        providerOptions: {},
       });
 
       expect(await server.calls[0].requestBody).toStrictEqual({
@@ -214,7 +189,7 @@ describe('GoogleVertexImageModel', () => {
         seed: 42,
         providerOptions: {
           vertex: {
-            temperature: 0.8,
+            addWatermark: false,
           },
         },
       });
@@ -225,7 +200,7 @@ describe('GoogleVertexImageModel', () => {
           sampleCount: 1,
           aspectRatio: '1:1',
           seed: 42,
-          temperature: 0.8,
+          addWatermark: false,
         },
       });
     });
@@ -302,7 +277,7 @@ describe('GoogleVertexImageModel', () => {
 
       const result = await model.doGenerate({
         prompt,
-        n: 1,
+        n: 2,
         size: undefined,
         aspectRatio: undefined,
         seed: undefined,
@@ -319,5 +294,36 @@ describe('GoogleVertexImageModel', () => {
       );
       expect(result.response.modelId).toBe('imagen-3.0-generate-001');
     });
+
+    it('should only pass valid provider options', async () => {
+      prepareJsonResponse();
+
+      await model.doGenerate({
+        prompt,
+        n: 2,
+        size: undefined,
+        aspectRatio: '16:9',
+        seed: undefined,
+        providerOptions: {
+          vertex: {
+            addWatermark: false,
+            negativePrompt: 'negative prompt',
+            personGeneration: 'allow_all',
+            foo: 'bar',
+          },
+        },
+      });
+
+      expect(await server.calls[0].requestBody).toStrictEqual({
+        instances: [{ prompt }],
+        parameters: {
+          sampleCount: 2,
+          addWatermark: false,
+          negativePrompt: 'negative prompt',
+          personGeneration: 'allow_all',
+          aspectRatio: '16:9',
+        },
+      });
+    });
   });
 });

packages/google-vertex/src/google-vertex-image-model.ts

@@ -3,6 +3,7 @@ import {
   Resolvable,
   combineHeaders,
   createJsonResponseHandler,
+  parseProviderOptions,
   postJsonToApi,
   resolve,
 } from '@ai-sdk/provider-utils';
@@ -65,13 +66,19 @@ export class GoogleVertexImageModel implements ImageModelV1 {
       });
     }
 
+    const vertexImageOptions = parseProviderOptions({
+      provider: 'vertex',
+      providerOptions,
+      schema: vertexImageProviderOptionsSchema,
+    });
+
     const body = {
       instances: [{ prompt }],
       parameters: {
         sampleCount: n,
         ...(aspectRatio != null ? { aspectRatio } : {}),
         ...(seed != null ? { seed } : {}),
-        ...(providerOptions.vertex ?? {}),
+        ...(vertexImageOptions ?? {}),
       },
     };
 
@@ -108,3 +115,23 @@ export class GoogleVertexImageModel implements ImageModelV1 {
 const vertexImageResponseSchema = z.object({
   predictions: z.array(z.object({ bytesBase64Encoded: z.string() })).nullish(),
 });
+
+const vertexImageProviderOptionsSchema = z.object({
+  negativePrompt: z.string().nullish(),
+  personGeneration: z
+    .enum(['dont_allow', 'allow_adult', 'allow_all'])
+    .nullish(),
+  safetySetting: z
+    .enum([
+      'block_low_and_above',
+      'block_medium_and_above',
+      'block_only_high',
+      'block_none',
+    ])
+    .nullish(),
+  addWatermark: z.boolean().nullish(),
+  storageUri: z.string().nullish(),
+});
+export type GoogleVertexImageProviderOptions = z.infer<
+  typeof vertexImageProviderOptionsSchema
+>;

packages/google-vertex/src/index.ts

@@ -1,3 +1,4 @@
+export type { GoogleVertexImageProviderOptions } from './google-vertex-image-model';
 export { createVertex, vertex } from './google-vertex-provider-node';
 export type {
   GoogleVertexProvider,

packages/google/src/google-generative-ai-language-model.test.ts

@@ -414,6 +414,39 @@ describe('doGenerate', () => {
     });
   });
 
+  it('should only pass valid provider options', async () => {
+    prepareJsonResponse({});
+
+    await model.doGenerate({
+      inputFormat: 'prompt',
+      mode: { type: 'regular' },
+      prompt: [
+        { role: 'system', content: 'test system instruction' },
+        { role: 'user', content: [{ type: 'text', text: 'Hello' }] },
+      ],
+      seed: 123,
+      temperature: 0.5,
+      providerMetadata: {
+        google: { foo: 'bar', responseModalities: ['TEXT', 'IMAGE'] },
+      },
+    });
+
+    expect(await server.calls[0].requestBody).toStrictEqual({
+      contents: [
+        {
+          role: 'user',
+          parts: [{ text: 'Hello' }],
+        },
+      ],
+      systemInstruction: { parts: [{ text: 'test system instruction' }] },
+      generationConfig: {
+        seed: 123,
+        temperature: 0.5,
+        responseModalities: ['TEXT', 'IMAGE'],
+      },
+    });
+  });
+
   it('should pass tools and toolChoice', async () => {
     prepareJsonResponse({});
 
@@ -1869,4 +1902,29 @@ describe('doStream', () => {
       'tool-calls',
     );
   });
+
+  it('should only pass valid provider options', async () => {
+    prepareStreamResponse({ content: [''] });
+
+    await model.doStream({
+      inputFormat: 'prompt',
+      mode: { type: 'regular' },
+      prompt: TEST_PROMPT,
+      providerMetadata: {
+        google: { foo: 'bar', responseModalities: ['TEXT', 'IMAGE'] },
+      },
+    });
+
+    expect(await server.calls[0].requestBody).toMatchObject({
+      contents: [
+        {
+          role: 'user',
+          parts: [{ text: 'Hello' }],
+        },
+      ],
+      generationConfig: {
+        responseModalities: ['TEXT', 'IMAGE'],
+      },
+    });
+  });
 });

packages/google/src/google-generative-ai-language-model.ts

@@ -88,9 +88,7 @@ export class GoogleGenerativeAILanguageModel implements LanguageModelV1 {
     const googleOptions = parseProviderOptions({
       provider: 'google',
       providerOptions: providerMetadata,
-      schema: z.object({
-        responseModalities: z.array(z.enum(['TEXT', 'IMAGE'])).nullish(),
-      }),
+      schema: googleGenerativeAIProviderOptionsSchema,
     });
 
     const generationConfig = {
@@ -623,3 +621,10 @@ const chunkSchema = z.object({
     })
     .nullish(),
 });
+
+const googleGenerativeAIProviderOptionsSchema = z.object({
+  responseModalities: z.array(z.enum(['TEXT', 'IMAGE'])).nullish(),
+});
+export type GoogleGenerativeAIProviderOptions = z.infer<
+  typeof googleGenerativeAIProviderOptionsSchema
+>;

packages/google/src/index.ts

@@ -1,6 +1,7 @@
-export { createGoogleGenerativeAI, google } from './google-provider';
 export type { GoogleErrorData } from './google-error';
+export type { GoogleGenerativeAIProviderOptions } from './google-generative-ai-language-model';
 export type { GoogleGenerativeAIProviderMetadata } from './google-generative-ai-prompt';
+export { createGoogleGenerativeAI, google } from './google-provider';
 export type {
   GoogleGenerativeAIProvider,
   GoogleGenerativeAIProviderSettings,