docs/pages/docs/ai-core/_meta.json

@@ -10,6 +10,8 @@
   "tools": "Tools and Tool Calling",
   "openai": "OpenAI Provider",
   "mistral": "Mistral Provider",
+  "google": "Google Provider",
+  "anthropic": "Anthropic Provider",
   "custom-provider": "Custom Providers",
   "generate-text": "generateText API",
   "stream-text": "streamText API",

docs/pages/docs/ai-core/anthropic.mdx

@@ -0,0 +1,50 @@
+---
+title: Anthropic Provider
+---
+
+import { Callout } from 'nextra-theme-docs';
+
+# Anthropic Provider
+
+<Callout>The Anthropic provider does not support tool calling yet.</Callout>
+
+The Anthropic provider contains language model support for the [Anthropic Messages API](https://docs.anthropic.com/claude/reference/messages_post).
+It creates language model objects that can be used with the `generateText` and `streamText`AI functions.
+
+## Provider Instance
+
+You can import `Anthropic` from `ai/anthropic` and initialize a provider instance with various settings:
+
+```ts
+import { Anthropic } from 'ai/anthropic';
+
+const anthropic = new Anthropic({
+  baseUrl: '', // optional base URL for proxies etc.
+  apiKey: '', // optional API key, default to env property ANTHROPIC_API_KEY
+});
+```
+
+The AI SDK also provides a shorthand `anthropic` import with a Anthropic provider instance that uses defaults:
+
+```ts
+import { anthropic } from 'ai/anthropic';
+```
+
+## Messages Models
+
+You can create models that call the [Anthropic Messages API](https://docs.anthropic.com/claude/reference/messages_post) using the `.messages()` factory method.
+The first argument is the model id, e.g. `claude-3-haiku-20240307`.
+Some models have multi-modal capabilities.
+
+```ts
+const model = anthropic.messages('claude-3-haiku-20240307');
+```
+
+Anthropic Messages` models support also some model specific settings that are not part of the [standard call settings](/docs/ai-core/settings).
+You can pass them as an options argument:
+
+```ts
+const model = anthropic.messages('claude-3-haiku-20240307', {
+  topK: 0.2,
+});
+```

docs/pages/docs/ai-core/google.mdx

@@ -0,0 +1,48 @@
+---
+title: Google Provider
+---
+
+import { Callout } from 'nextra-theme-docs';
+
+# Google Provider
+
+The Google provider contains language model support for the [Google Generative AI](https://ai.google/discover/generativeai/) APIs.
+It creates language model objects that can be used with the `generateText`, `streamText`, `generateObject`, and `streamObject` AI functions.
+
+## Provider Instance
+
+You can import `Google` from `ai/google` and initialize a provider instance with various settings:
+
+```ts
+import { Google } from 'ai/google';
+
+const google = new Google({
+  baseUrl: '', // optional base URL for proxies etc.
+  apiKey: '', // optional API key, default to env property GOOGLE_GENERATIVE_AI_API_KEY
+});
+```
+
+The AI SDK also provides a shorthand `google` import with a Google provider instance that uses defaults:
+
+```ts
+import { google } from 'ai/google';
+```
+
+## Generative AI Models
+
+You can create models that call the [Google Generative AI API](https://ai.google.dev/api/rest) using the `.generativeAI()` factory method.
+The first argument is the model id, e.g. `models/gemini-pro`.
+The models support tool calls and some have multi-modal capabilities.
+
+```ts
+const model = google.generativeAI('models/gemini-pro');
+```
+
+Google Generative AI models support also some model specific settings that are not part of the [standard call settings](/docs/ai-core/settings).
+You can pass them as an options argument:
+
+```ts
+const model = google.generativeAI('models/gemini-pro', {
+  topK: 0.2,
+});
+```

docs/pages/docs/ai-core/index.mdx

@@ -74,6 +74,8 @@ The AI SDK contains the following providers:
 
 - [OpenAI Provider](/docs/ai-core/openai) (`ai/openai`)
 - [Mistral Provider](/docs/ai-core/mistral) (`ai/mistral`)
+- [Google Provider](/docs/ai-core/google) (`ai/google`)
+- [Anthropic Provider](/docs/ai-core/anthropic) (`ai/anthropic`)
 
 The AI SDK also provides a [language model specification](https://github.com/vercel/ai/tree/main/packages/core/spec/language-model/v1) that you can use to implement [custom providers](/docs/ai-core/custom-provider).
 

examples/ai-core/.env.example

@@ -1,3 +1,4 @@
+ANTHROPIC_API_KEY=""
 OPENAI_API_KEY=""
 MISTRAL_API_KEY=""
-GOOGLE_GENERATIVE_AI_API_KEY=""
+GOOGLE_GENERATIVE_AI_API_KEY=""

examples/ai-core/src/generate-text/anthropic-multimodal.ts

@@ -0,0 +1,26 @@
+import { experimental_generateText } from 'ai';
+import { anthropic } from 'ai/anthropic';
+import dotenv from 'dotenv';
+import fs from 'node:fs';
+
+dotenv.config();
+
+async function main() {
+  const result = await experimental_generateText({
+    model: anthropic.messages('claude-3-haiku-20240307'),
+    maxTokens: 512,
+    messages: [
+      {
+        role: 'user',
+        content: [
+          { type: 'text', text: 'Describe the image in detail.' },
+          { type: 'image', image: fs.readFileSync('./data/comic-cat.png') },
+        ],
+      },
+    ],
+  });
+
+  console.log(result.text);
+}
+
+main();

examples/ai-core/src/generate-text/anthropic.ts

@@ -0,0 +1,19 @@
+import { experimental_generateText } from 'ai';
+import { anthropic } from 'ai/anthropic';
+import dotenv from 'dotenv';
+
+dotenv.config();
+
+async function main() {
+  const result = await experimental_generateText({
+    model: anthropic.messages('claude-3-haiku-20240307'),
+    prompt: 'Invent a new holiday and describe its traditions.',
+  });
+
+  console.log(result.text);
+  console.log();
+  console.log('Token usage:', result.usage);
+  console.log('Finish reason:', result.finishReason);
+}
+
+main();

examples/ai-core/src/stream-text/anthropic-multimodal.ts

@@ -0,0 +1,28 @@
+import { experimental_streamText } from 'ai';
+import { anthropic } from 'ai/anthropic';
+import dotenv from 'dotenv';
+import fs from 'node:fs';
+
+dotenv.config();
+
+async function main() {
+  const result = await experimental_streamText({
+    model: anthropic.messages('claude-3-haiku-20240307'),
+    maxTokens: 512,
+    messages: [
+      {
+        role: 'user',
+        content: [
+          { type: 'text', text: 'Describe the image in detail.' },
+          { type: 'image', image: fs.readFileSync('./data/comic-cat.png') },
+        ],
+      },
+    ],
+  });
+
+  for await (const textPart of result.textStream) {
+    process.stdout.write(textPart);
+  }
+}
+
+main();

examples/ai-core/src/stream-text/anthropic.ts

@@ -0,0 +1,18 @@
+import { experimental_streamText } from 'ai';
+import { anthropic } from 'ai/anthropic';
+import dotenv from 'dotenv';
+
+dotenv.config();
+
+async function main() {
+  const result = await experimental_streamText({
+    model: anthropic.messages('claude-3-haiku-20240307'),
+    prompt: 'Invent a new holiday and describe its traditions.',
+  });
+
+  for await (const textPart of result.textStream) {
+    process.stdout.write(textPart);
+  }
+}
+
+main();

packages/core/CHANGELOG.md

@@ -1,5 +1,12 @@
 # ai
 
+## 3.0.18
+
+### Patch Changes
+
+- 63d587e: Add Anthropic provider for ai/core functions (no tool calling).
+- 63d587e: Add automatic mime type detection for images in ai/core prompts.
+
 ## 3.0.17
 
 ### Patch Changes

packages/core/anthropic/anthropic-error.ts

@@ -0,0 +1,17 @@
+import { z } from 'zod';
+import { createJsonErrorResponseHandler } from '../spec';
+
+const anthropicErrorDataSchema = z.object({
+  type: z.literal('error'),
+  error: z.object({
+    type: z.string(),
+    message: z.string(),
+  }),
+});
+
+export type AnthropicErrorData = z.infer<typeof anthropicErrorDataSchema>;
+
+export const anthropicFailedResponseHandler = createJsonErrorResponseHandler({
+  errorSchema: anthropicErrorDataSchema,
+  errorToMessage: data => data.error.message,
+});

packages/core/anthropic/anthropic-facade.ts

@@ -0,0 +1,58 @@
+import { generateId, loadApiKey } from '../spec';
+import { AnthropicMessagesLanguageModel } from './anthropic-messages-language-model';
+import {
+  AnthropicMessagesModelId,
+  AnthropicMessagesSettings,
+} from './anthropic-messages-settings';
+
+/**
+ * Anthropic provider.
+ */
+export class Anthropic {
+  readonly baseUrl?: string;
+  readonly apiKey?: string;
+
+  private readonly generateId: () => string;
+
+  constructor(
+    options: {
+      baseUrl?: string;
+      apiKey?: string;
+      generateId?: () => string;
+    } = {},
+  ) {
+    this.baseUrl = options.baseUrl;
+    this.apiKey = options.apiKey;
+    this.generateId = options.generateId ?? generateId;
+  }
+
+  private get baseConfig() {
+    return {
+      baseUrl: this.baseUrl ?? 'https://api.anthropic.com/v1',
+      headers: () => ({
+        'anthropic-version': '2023-06-01',
+        'x-api-key': loadApiKey({
+          apiKey: this.apiKey,
+          environmentVariableName: 'ANTHROPIC_API_KEY',
+          description: 'Anthropic',
+        }),
+      }),
+    };
+  }
+
+  messages(
+    modelId: AnthropicMessagesModelId,
+    settings: AnthropicMessagesSettings = {},
+  ) {
+    return new AnthropicMessagesLanguageModel(modelId, settings, {
+      provider: 'anthropic.messages',
+      ...this.baseConfig,
+      generateId: this.generateId,
+    });
+  }
+}
+
+/**
+ * Default Anthropic provider instance.
+ */
+export const anthropic = new Anthropic();