Adding Generative AI capabilities to your data in Switzerland with Amazon Bedrock

Today, we are announcing the availability of Amazon Bedrock in the Europe (Zurich) region. Customers can now access Amazon Bedrock via an API endpoint in Switzerland and have the option to process their data in country.

In this post, we are providing an overview of the new Amazon Bedrock capabilities launching in the Europe (Zurich) region, including the key features and models available. We will also cover how customers can get started using Amazon Bedrock in this new region.

Swiss Post uses Amazon Bedrock to build generative AI-based applications to increase productivity and efficiency. For example, they built LLM4Post — based on Amazon Bedrock — as a playground to allow the company´s internal data scientists to quickly validate use cases with stakeholders.
“We like Amazon Bedrock because it is easy to use and integrate into our existing AWS infrastructure landscape as well as the fast pace of new features being implemented,” said Philippe Goetschmann, Lead ML Engineer at Swiss Post.

What is Amazon Bedrock?

Generative AI is one of the most transformative technologies of our generation, addressing some of humanity’s most pressing challenges, augmenting human performance, and maximizing productivity. AWS democratizes access to generative AI technology by enabling companies of all sizes and industries. We empower developers of all skills levels to participate in this transformation and to grow their businesses. Amazon Bedrock provides a choice of foundation models from leading AI providers via a single API, enabling companies to build and scale generative AI applications easily.
At AWS, security and privacy are top priority, and these principles are at the core of Amazon Bedrock. With Amazon Bedrock we do not store user input or model output, nor do we share this information with third-party model providers. For more details, please refer to the blog post on “Data Protection and Responsible use of Amazon Bedrock in Switzerland“.

What is available in the AWS Europe (Zurich) Region?

As of today, AWS is offering Amazon Bedrock in the Europe (Zurich) region. This launch includes the following features and models.

Foundational models available at launch:

Amazon Titan: Text Embeddings V2
Anthropic Claude: Claude 3.5 Sonnet, Claude 3.0 Haiku

AWS continues to deploy additional features and models in response to high customer demand.

To list all available models see Model support by AWS Region, or using the AWS CLI, run the following command:

% aws bedrock list-foundation-models --query "modelSummaries[*].modelId" --region eu-central-2
[
    "amazon.titan-embed-text-v2:0:8k",
    "amazon.titan-embed-text-v2:0",
    "anthropic.claude-3-haiku-20240307-v1:0:48k",
    "anthropic.claude-3-haiku-20240307-v1:0:200k",
    "anthropic.claude-3-haiku-20240307-v1:0",
    "anthropic.claude-3-5-sonnet-20240620-v1:0:18k",
    "anthropic.claude-3-5-sonnet-20240620-v1:0:51k",
    "anthropic.claude-3-5-sonnet-20240620-v1:0:200k",
    "anthropic.claude-3-5-sonnet-20240620-v1:0"
]

Additional functionality available with Amazon Bedrock in the region:

VPC Endpoint – create a private connection between your VPC and Amazon Bedrock
Guardrails for Amazon Bedrock – implement safeguards to filter harmful content, redact sensitive information (PII), safeguard against prompt attacks, or detect hallucinations
Cross-region inference (available soon) – optional feature to manage traffic bursts by utilizing compute across different AWS Regions
Agents for Amazon Bedrock (available soon) – orchestrate and execute multistep tasks
Knowledge Bases for Amazon Bedrock (available soon) – fully managed support for end-to-end Retrieval Augmented Generation (RAG) workflow
Model evaluation (available soon) – select the foundation model that gives you the best results for your particular use case
Prompt flows (available soon) – quickly create workflows by leveraging generative AI

Create, test, and deploy workflows through an intuitive visual builder!
Prompt Flows: A visual builder to link foundation models (FMs), prompts, and AWS services together.

How to get started?

Enable AWS Europe(Zurich) and foundational models

First, you need an AWS account and you need to enable the AWS Europe (Zurich) region in that account. Then, you need to enable the foundational models you intend to use in Amazon Bedrock. Ensure you are accessing Amazon Bedrock in the AWS Europe/Zurich region.

Getting started with playgrounds (AWS accounts)

The Amazon Bedrock playgrounds provide a console environment to experiment with running inference on different models and with different configurations, before deciding to use them in an application. In the console, access the playgrounds by choosing Playgrounds in the left navigation pane.

Bedrock Playgrounds

Playground for end-users

End users can start learning generative AI fundamentals, using Amazon PartyRock. Making learning easy with a hands-on, code-free app builder. Experiment with prompt engineering techniques, review generated responses, and develop intuition for generative AI while creating and exploring fun apps. Amazon PartyRock leverages Amazon Bedrock in the back. Here is a playlist with some PartyRock apps.

Accessing Amazon Bedrock programmatically and generating prompts

AWS software development kits (SDKs) are available for programming languages such as C++, Go, Java, Javascript, Kotlin, .Net, PHP, Python, Ruby, Rust, ABAP and Swift. Each SDK provides an API, code examples, and documentation that enables developers to build applications in their preferred language. To get you started quickly, we recommend the following two videos:
Integrating Generative AI Models with Amazon Bedrock
Agents Tools & Function Calling with Amazon Bedrock

We encourage you to look specifically into the Amazon Bedrock Converse API, which provides a consistent way to access large language models (LLMs) using Amazon Bedrock. Previously, the InvokeModel API required different JSON structures for requests and responses across model providers. The Converse API now allows a single format for requests and responses across all large language models on Amazon Bedrock. Check “Getting started with the Amazon Bedrock Converse API” for more details. Amazon Bedrock Recipes provides a collection of resources to help developers to use and learn about the features of Amazon Bedrock.

Hands-On workshops and prompt engineering tools

At AWS, we are builders, and we love to engage with our customers hands-on. To that end, we create workshops that provide step-by-step guides on how to get things done with Amazon Bedrock. We encourage you to run through these workshops yourself to get started with Amazon Bedrock. You can find a wide range of workshops for Amazon Bedrock covering the above use-cases and more.

Additionally, we want you to get started quickly to create powerful prompts for the models from Anthropic. Prompt Generator is a prompt engineering tool designed to solve the “blank page problem” and gives a starting point for iteration. You can access the Prompt Generator on our PartyRock app.

Summary

With Amazon Bedrock, we are democratizing access to generative AI technology and making it as easy as possible to infuse generative AI into your applications. The launch of Amazon Bedrock in Switzerland addresses a customer need to have an API endpoint and data processing capabilities within the country. We have built Amazon Bedrock with security, privacy and responsibility in mind. We commit to not storing or using any data that is sent or received from the foundational models For more details on these commitments, please refer to the accompanying blog post. We are not stopping here. We encourage our customers to share their interest in additional features and models by contacting your AWS sales representatives.

AWS in Switzerland and Austria (Alps)