From 2ec91c406bce1e42613d8d2f153fbe1b159654a0 Mon Sep 17 00:00:00 2001
From: Fumito Ito <weathercook@gmail.com>
Date: Fri, 4 Oct 2024 12:26:59 +0900
Subject: [PATCH] Update README.md

---
 README.md | 110 +++---------------------------------------------------
 1 file changed, 5 insertions(+), 105 deletions(-)

diff --git a/README.md b/README.md
index 34c5912..fd189da 100644
--- a/README.md
+++ b/README.md
@@ -106,113 +106,13 @@ let result = try await Anthropic(apiKey: "your_claude_api_key")
     )
 ```
 
-## Amazon Web Services Bedrock
+## Extensions
 
-This library provides support for the [Anthropic Bedrock API](https://aws.amazon.com/bedrock/claude/) through a separate package.
+By introducing an extension Swift package, it is possible to access the Anthropic Claude API through AWS Bedrock and Vertex AI. The supported services are as follows:
 
-```swift
-let package = Package(
-    name: "MyPackage",
-    products: [...],
-    targets: [
-        .target(
-            "YouAppModule",
-            dependencies: [
-                .product(name: "AnthropicSwiftSDK-Bedrock", package: "AnthropicSwiftSDK")
-            ]
-        )
-    ],
-    dependencies: [
-        .package(url: "https://github.com/fumito-ito/AnthropicSwiftSDK.git", .upToNextMajor(from: "0.5.0"))
-    ]
-)
-```
-
-To create an `AnthropicBedrockClient` from a `BedrockRuntimeClient` with a `Model` to access Claude on Bedrock.
-The API usage is the same as the normal AnthropicClient.
-
-```swift
-let client = try BedrockRuntimeClient(region: "us-west-2")
-let anthropic = client.useAnthropic()
-
-let response = try await anthropic.messages.createMessage(Message(role: .user, content: [.text("This is test text")]), maxTokens: 1024)
-for content in response.content {
-    switch content {
-    case .text(let text):
-        print(text)
-    case .image(let imageContent):
-        // handle base64 encoded image content
-    }
-}
-```
-
-Of course, `Streaming Message API` works in the same way.
-
-```swift
-let client = try BedrockRuntimeClient(region: "us-west-2")
-let anthropic = client.useAnthropic()
-
-let stream = try await anthropic.messages.streamMessage([Message(role: .user, content: [.text("This is test text")])], maxTokens: 1024)
-for try await chunk in stream {
-    switch chunk.type {
-    case .messageStart:
-        // handle message start object with casting chunk into `StreamingMessageStartResponse`
-    }
-}
-```
-
-## Google Vertex AI
-
-This library provides support for the [Anthropic Vertex AI](https://cloud.google.com/blog/products/ai-machine-learning/announcing-anthropics-claude-3-models-in-google-cloud-vertex-ai?hl=en) through a separate package.
-
-```swift
-let package = Package(
-    name: "MyPackage",
-    products: [...],
-    targets: [
-        .target(
-            "YouAppModule",
-            dependencies: [
-                .product(name: "AnthropicSwiftSDK-VertexAI", package: "AnthropicSwiftSDK")
-            ]
-        )
-    ],
-    dependencies: [
-        .package(url: "https://github.com/fumito-ito/AnthropicSwiftSDK.git", .upToNextMajor(from: "0.5.0"))
-    ]
-)
-```
-
-To create an `AnthropicVertexAIClient` with a `Model` to access Claude on VertexAI.
-The API usage is the same as the normal AnthropicClient.
-
-```swift
-let anthropic = AnthropicVertexAIClient(projectId: "your-project-id", accessToken: "access-token-for-vertexai", region = .usCentral1)
-
-let response = try await anthropic.messages.createMessage(Message(role: .user, content: [.text("This is test text")]), maxTokens: 1024)
-for content in response.content {
-    switch content {
-    case .text(let text):
-        print(text)
-    case .image(let imageContent):
-        // handle base64 encoded image content
-    }
-}
-```
-
-Of course, `Streaming Message API` works in the same way.
-
-```swift
-let anthropic = AnthropicVertexAIClient(projectId: "your-project-id", accessToken: "access-token-for-vertexai", region = .usCentral1)
-
-let stream = try await anthropic.messages.streamMessage([Message(role: .user, content: [.text("This is test text")])], maxTokens: 1024)
-for try await chunk in stream {
-    switch chunk.type {
-    case .messageStart:
-        // handle message start object with casting chunk into `StreamingMessageStartResponse`
-    }
-}
-```
+- [Amazon Web Services Bedrock](https://github.com/fumito-ito/AnthropicSwiftSDK-Bedrock)
+- VertexAI
+    - T.B.D.
 
 ## Contributing