|
Replace Might 22, 2024: This submit has been up to date to incorporate a hyperlink to the Amazon Bedrock Studio bootstrapper script that automates the creation of the permissions boundary, service function, and provisioning function required to create workspaces in Amazon Bedrock Studio.
Replace Might 7, 2024: Up to date screenshots on this submit to mirror current updates to the UI.
Immediately, we’re introducing Amazon Bedrock Studio, a brand new web-based generative synthetic intelligence (generative AI) growth expertise, in public preview. Amazon Bedrock Studio accelerates the event of generative AI functions by offering a speedy prototyping atmosphere with key Amazon Bedrock options, together with Data Bases, Brokers, and Guardrails.
As a developer, now you can use your organization’s single sign-on credentials to sign up to Bedrock Studio and begin experimenting. You possibly can construct functions utilizing a big selection of prime performing fashions, consider, and share your generative AI apps inside Bedrock Studio. The consumer interface guides you thru varied steps to assist enhance a mannequin’s responses. You possibly can experiment with mannequin settings, and securely combine your organization knowledge sources, instruments, and APIs, and set guardrails. You possibly can collaborate with workforce members to ideate, experiment, and refine your generative AI functions—all with out requiring superior machine studying (ML) experience or AWS Administration Console entry.
As an Amazon Internet Companies (AWS) administrator, you could be assured that builders will solely have entry to the options offered by Bedrock Studio, and gained’t have broader entry to AWS infrastructure and providers.
Now, let me present you methods to get began with Amazon Bedrock Studio.
Get began with Amazon Bedrock Studio
As an AWS administrator, you first must create an Amazon Bedrock Studio workspace, then choose and add customers you wish to give entry to the workspace. As soon as the workspace is created, you may share the workspace URL with the respective customers. Customers with entry privileges can sign up to the workspace utilizing single sign-on, create tasks inside their workspace, and begin constructing generative AI functions.
Create Amazon Bedrock Studio workspace
Navigate to the Amazon Bedrock console and select Bedrock Studio on the underside left pane.
Earlier than making a workspace, you could configure and safe the only sign-on integration together with your identification supplier (IdP) utilizing the AWS IAM Id Middle. For detailed directions on methods to configure varied IdPs, similar to AWS Listing Service for Microsoft Energetic Listing, Microsoft Entra ID, or Okta, take a look at the AWS IAM Id Middle Consumer Information. For this demo, I configured consumer entry with the default IAM Id Middle listing.
Subsequent, select Create workspace, enter your workspace particulars, and create any required AWS Id and Entry Administration (IAM) roles. As a substitute for following the directions offered within the Amazon Bedrock Consumer Information, you should use the Amazon Bedrock Studio bootstrapper script to create the permissions boundary, service function, and provisioning function. The script contains directions on methods to customise the configuration if wanted.
If you’d like, it’s also possible to choose default generative AI fashions and embedding fashions for the workspace. When you’re performed, select Create.
Subsequent, choose the created workspace.
Then, select Consumer administration and Add customers or teams to pick out the customers you wish to give entry to this workspace.
Again within the Overview tab, now you can copy the Bedrock Studio URL and share it together with your customers.
Construct generative AI functions utilizing Amazon Bedrock Studio
As a builder, now you can navigate to the offered Bedrock Studio URL and sign up together with your single sign-on consumer credentials. Welcome to Amazon Bedrock Studio! Let me present you ways to select from trade main FMs, deliver your individual knowledge, use features to make API calls, and safeguard your functions utilizing guardrails.
Select from a number of trade main FMs
By selecting Discover, you can begin choosing accessible FMs and discover the fashions utilizing pure language prompts.
When you select Construct, you can begin constructing generative AI functions in a playground mode, experiment with mannequin configurations, iterate on system prompts to outline the conduct of your software, and prototype new options.
Deliver your individual knowledge
With Bedrock Studio, you may securely deliver your individual knowledge to customise your software by offering a single file or by choosing a data base created in Amazon Bedrock.
Use features to make API calls and make mannequin responses extra related
A operate name permits the FM to dynamically entry and incorporate exterior knowledge or capabilities when responding to a immediate. The mannequin determines which operate it must name based mostly on an OpenAPI schema that you simply present.
Capabilities allow a mannequin to incorporate info in its response that it doesn’t have direct entry to or prior data of. For instance, a operate may permit the mannequin to retrieve and embody the present climate circumstances in its response, though the mannequin itself doesn’t have that info saved.
Safeguard your functions utilizing Guardrails for Amazon Bedrock
You possibly can create guardrails to advertise protected interactions between customers and your generative AI functions by implementing safeguards personalized to your use circumstances and accountable AI insurance policies.
If you create functions in Amazon Bedrock Studio, the corresponding managed sources similar to data bases, brokers, and guardrails are mechanically deployed in your AWS account. You need to use the Amazon Bedrock API to entry these sources in downstream functions.
Right here’s a brief demo video of Amazon Bedrock Studio created by my colleague Banjo Obayomi.
Be a part of the preview
Amazon Bedrock Studio is accessible as we speak in public preview in AWS Areas US East (N. Virginia) and US West (Oregon). To be taught extra, go to the Amazon Bedrock Studio web page and Consumer Information.
Give Amazon Bedrock Studio a strive as we speak and tell us what you assume! Ship suggestions to AWS re:Put up for Amazon Bedrock or via your regular AWS contacts, and have interaction with the generative AI builder neighborhood at neighborhood.aws.
— Antje
