Anthropic Library

I am gonna give you a little preview of the new version of the Anthropic Library. One of the cool new features that it has in it is a data agent component that makes it very easy to prepare data and interact with your data before you have the user interface in place to do the real work. So I have a little demonstration app here. It has the Anthropic library as a dependency. We have two business objects in here. Applicant with last name, first name, email. Applicant can have many addresses. An address belongs to an applicant has typical properties. It also has an address type, which is a base type with legal and mailing, um, valid values. Given all of that, that's it for the business objects. I have a single page in this application, which is very simple. It's just a flex column, a row with a header and some cost tracking. So you'll see that in action. And then we've placed in the anthropic data agent component, uh, and set it to take up the rest of the page. So let's see that in action. So here we are on the page. We haven't started yet, so we're not spending any anthropic uh, token, uh, yet. So you can ask the agent what it can do and it will give you an answer that describes the different business objects that it knows about and what it can do to those objects. Then the fun starts when you say something like, create 10 synthetic applicants for testing I, Okay, it's followed the schema and created the appropriate properties, put nice fake values in them. Now we can interact with this list as well. We can say sort by last name and it will actually use the loop back API to prepare the correct filter for you and order those applicants by name. In addition to showing a tabular form on the right, we can uh, use an an individual page. So let's say show me Sarah and it will open up the detail page for Sarah that shows the fields on that applicant as well as the addresses. There are none, so we can say add to synthetic addresses and it will make them up. Notice that it knew that the types are legal and mailing, and so it picked that automatically. That was based on the schema and Apex designer. And now we can e make even more general comments, like do the same for all other applicants. It knows there are nine other ones because it's listed them before it's creating them. In this case it decided to do them one by one and in some cases it will end up making them all in one shot. And then you can do something like show me a different applicant And it just switched to some other applicant. It also knows information on the left, like the idea of the applicant and the summary of the information. So you can see that it's really quite, uh, straightforward. It's entirely driven by the schema of the business objects and the standard CRUD operations that are available from the loop back, API in Apex designer and on the right. What we're going to do as we go forward is extend not just this sort of generic table and fields plus relationships presentation, but also allow you to set, uh, define sections that are custom presentations for certain business objects. So that gives you an idea what this data agent capability is all about, and hopefully that gets your wheels turning about how you can leverage anthropic in your Apex Designer applications to create agent-based applications.