Today, we are announcing the preview launch of Amazon Bedrock Prompt Management and Prompt Flows. Amazon Bedrock Prompt Management simplifies the creation, evaluation, versioning, and sharing of prompts to help developers and prompt engineers get the best responses from foundation models for their use cases. Developers can use the Prompt Builder to experiment with multiple FMs, model configurations, and prompt messages. They can test and compare prompts in-place using the Prompt Builder, without the need of any deployment. To share the prompt for use in downstream applications, they can simply create a version and make an API call to retrieve the prompt. In addition, Bedrock Prompt Flows accelerates the creation, testing, and deployment of workflows through an intuitive visual builder. Developers can use the visual builder to drag and drop different components such as prompts, Knowledge Bases, and Lambda functions to automate a workflow.

Knowledge Bases for Amazon Bedrock is a fully managed Retrieval-Augmented Generation (RAG) capability that allows you to connect foundation models (FMs) to internal company data sources to deliver relevant and accurate responses. Chunking allows processing long documents by breaking them into smaller chunks, enabling accurate knowledge retrieval from a user’s question. Today, we are launching advanced chunking options. The first is custom chunking. With this, customers can write their own chunking code as a Lambda function, and even use off the shelf components from frameworks like LangChain and LlamaIndex. Additionally, we are launching built-in chunking options such as semantic and hierarchical chunking.

Additionally, customers can enable smart parsing to extract information from more complex data such as tables. This capability uses Amazon Bedrock foundation models to parse tabular content in file formats such as PDF to improve retrieval accuracy. You can customize parsing prompts to extract data in the format of your choice. Knowledge Bases now also supports query reformulation. This capability breaks down queries into simpler sub-queries, retrieves relevant information for each, and combines the results into a final comprehensive answer. With these new accuracy improvements for chunking, parsing, and advanced query handling, Knowledge Bases empowers users to build highly accurate and relevant knowledge resources suited for enterprise use cases.

These capabilities are supported in the all AWS Regions where Knowledge Bases is available. To learn more about these features and how to get started, refer to the Knowledge Bases for Amazon Bedrock documentation and visit the Amazon Bedrock console .

Knowledge Bases for Amazon Bedrock is a fully managed Retrieval-Augmented Generation (RAG) capability that allows you to connect foundation models (FMs) to internal company data sources to deliver relevant and accurate responses. Today, we are launching a new feature that allows customers to securely ingest data from various sources into their knowledge bases. Knowledge Bases now supports the web data source allowing you to index public web pages. Secondly, Knowledge Bases now supports three additional data connectors including Atlassian Confluence, Microsoft SharePoint, and Salesforce. You can connect directly to these data sources to build your RAG applications. These new capabilities reduce the time and cost associated with data movement, while ensuring that the knowledge bases stays up-to-date with the latest changes in the connected data sources.

Customers can set up these new data sources through the AWS Management Console for Amazon Bedrock or the CreateDataSource API. To get started, visit the Knowledge Bases documentation.

This capability is supported in the all AWS Regions where Knowledge Bases is available. To learn more about these features and how to get started, refer to the Knowledge Bases for Amazon Bedrock documentation and visit the Amazon Bedrock console .

Amazon Web Services, Inc. (AWS) today announced a new code interpretation capability on Agents for Amazon Bedrock. Code interpretation allows agents to dynamically generate and execute code snippets within a secure sandboxed environment, extending the capabilities of Agents for complex use cases such as data analysis, data visualization, and optimization problems.

This new capability allows developers to move beyond the predefined capabilities of the large language model (LLM) and tackle more complex, data-driven use cases. Agents can now generate and execute code, process files with diverse data types and formatting, and even generate graphs to enhance the user experience. Also, the iterative code execution capabilities allow Agents to work through challenging data science problems, giving them the ability to orchestrate increasingly complex tasks.

Code interpretation is currently available in the Northern Virginia, Oregon, Europe (Frankfurt) AWS regions.

Learn more here .

Amazon Web Services, Inc. (AWS) today announced Agents for Amazon Bedrock can retain memory across multiple interactions over time, allowing developers to build generative AI applications that seamlessly adapt to user context and preferences, enhancing personalized experiences and automating complex business processes more efficiently.

By retaining memory AI assistants remember historical knowledge and learn from user interactions over time. For example, if a user is booking a flight, the application can remember the user’s travel preferences for future bookings. This capability is crucial for complex multi-step tasks like insurance claims processing, where continuity and context retention significantly improve the user experience.

Memory retention is available in all AWS Regions where Claude 3 Sonnet and Haiku models support Agents for Amazon Bedrock.

Learn more about memory retention on Agents for Amazon Bedrock here .

Today, AWS announces the general availability of customized Amazon Q Developer inline code recommendations. You can now securely connect Amazon Q Developer to your private code bases and generate more precise suggestions by including your organization’s internal APIs, libraries, classes, methods, and best practices. In preview, you can also use Amazon Q Developer chat in the IDE to ask questions about how your internal code base is structured, where and how certain functions or libraries are used, or what specific functions, methods, or APIs do. With these capabilities, Amazon Q Developer can save builders hours typically spent examining previously written code or internal documentation to understand how to use internal APIs, libraries, and more. 

 

To get started, you first need to securely connect your organization’s private repositories to Amazon Q Developer in the AWS Management Console. Amazon Q Developer administrators can select which repositories to use to customize recommendations, applying strict access control. Your administrators can decide which customization to activate, and they can manage access to a private customization from the console so only specific developers have access. Each customization is isolated from other customers, and none of the customizations built with these new capabilities will be used to train the foundation models underlying Amazon Q Developer.

 

Customized code recommendations and chat in the IDE are available as part of the Amazon Q Developer Pro subscription. To learn more about pricing, visit Amazon Q Developer Pricing . To learn more about these capabilities, see Amazon Q Developer  or read the announcement blog post.

Guardrails for Amazon Bedrock enables customers to implement safeguards based on their application requirements and responsible AI policies. Today, guardrails adds contextual grounding checks and introduces a new ApplyGuardrail API to build trustworthy generative AI applications using any foundation model (FM).

Customers rely on the inherent capabilities of the FMs to generate grounded (credible) responses that are based on company’s source data. However, FMs can conflate multiple pieces of information, producing incorrect or new information – impacting the reliability of the application. With contextual grounding checks, Guardrails can now detect hallucinations in model responses for RAG (retrieval-augmented generation) and conversational applications. This safeguard helps detect and filter responses that are factually incorrect based on a reference source, and are irrelevant to the users’ query. Customers can configure confidence thresholds to filter responses with low confidence of grounding or relevance.

In addition, to support choice of safeguarding applications using different FMs, Guardrails now supports an ApplyGuardrail API to evaluate user inputs and model responses for any custom and third-party FM, in addition to FMs already supported in Amazon Bedrock. The ApplyGuardrail API now enables centralized safety and governance for all your generative AI applications.

Guardrails is the only offering from a major cloud provider to provide safety, privacy, and truthfulness protections in a single solution. Contextual grounding check and ApplyGuardrail API are supported in all AWS regions where Guardrails for Amazon Bedrock is supported.

To learn more about Guardrails for Amazon Bedrock, visit the feature page and read the news blog .
 

Today, AWS Backup announces support for Amazon EBS Snapshots Archive in backup policies, allowing customers to automatically move Amazon EBS Snapshots created by AWS Backup to Amazon EBS Snapshots Archive at the AWS Organizations level. Amazon EBS Snapshots Archive is low-cost, long-term storage tier meant for your rarely-accessed snapshots that do not need frequent retrieval. You can now use your Organizations’ management account to set an Amazon EBS Snapshots Archival policy across accounts.

To get started, create a new or edit an existing AWS Backup policy from your AWS Organizations’ management account. You can use AWS Backup policies to transition your Amazon EBS Snapshots to Amazon EBS Snapshots Archive and manage their lifecycle, alongside AWS Backup’s other supported resources. Amazon EBS Snapshots are incremental, storing only the changes since the last snapshot and making them cost effective for daily and weekly backups that need to be accessed frequently. You may also have Amazon EBS snapshots that you only need to access every few months, retaining them for long-term regulatory requirements. For these long-term snapshots, you can now transition your Amazon EBS snapshots managed by AWS Backup to Amazon EBS Snapshots Archive tier to store full snapshots at lower costs.

AWS Backup support for Amazon EBS Snapshots Archive in backup policies is available in all commercial and AWS GovCloud (US) Regions, where AWS Backup, AWS Backup policies and EBS Snapshots Archive are available. You can get started by using the AWS Organizations API, or CLI. For more information, visit our documentation.

Amazon SageMaker, a fully managed machine learning service, announces the general availability of Amazon Q Developer in SageMaker Studio. SageMaker Studio customers now get generative AI assistance powered by Q Developer right within their JupyterLab Integrated Development Environment (IDE). With Q Developer, data scientists and ML engineers can access expert guidance on SageMaker features, code generation, and troubleshooting. This allows for more productivity by eliminating the need for tedious online searches and documentation review, and ensuring more time delivering differentiated business value.

Data scientists and ML engineers using JupyterLab in SageMaker Studio can kick off their model development lifecycle with Amazon Q Developer. They can use the chat capability to discover and learn how to leverage SageMaker features for their use case without having to sift through extensive documentation. As well, users can generate code tailored to their needs and jump-start the development process. Further, they can use Q Developer to get in-line code suggestions and conversational assistance to edit, explain, and document their code in JupyterLab. Users can also leverage Q Developer to receive step by step guidance for troubleshooting when running into errors. With the introduction of Q Developer, users can leverage generative AI assistance within their JupyterLab environment. This integration empowers data scientists and ML engineers to accelerate their workflow, enhance productivity, and deliver ML models more efficiently, streamlining the machine learning development process.

This feature is available in all commercial AWS regions where SageMaker Studio is available.

For additional details, see our product page and documentation .

Starting today, customers can use Amazon Cognito in Asia Pacific (Hong Kong) Region. Cognito makes it easy to add authentication, authorization, and user management to your web and mobile apps. The service scales to millions of users and supports sign-in with social identity providers such as Apple, Facebook, Google, and Amazon, and enterprise identity providers via standards such as SAML 2.0 and OpenID Connect.

With the addition of this region, Amazon Cognito is now available in 29 AWS Regions globally. For a list of regions where Amazon Cognito is available, see the AWS Region Table . To learn more about Amazon Cognito, visit the product documentation page . To get started, visit the Amazon Cognito home page .