With a full-stack AI Suite, you can build full-stack generative AI capabilities without cloud architecture or machine learning experience. You don't need to worry about configuring resources or providing secure front-end access, and it's serverless, allowing for fast iteration, so you only pay for what you use.
In case you haven't heard, at AWS Amplify, we're all on typescript. With Amplify Gen 2, every part of the cloud back end of the application is defined in TypeScript. Need to authorize the backend? A typescript. The data back end? A typescript. Storage back end? A typescript. Generative full stack with AI? A typescript.
In this blog post, we'll review what's in the Amplify AI suite and how it simplifies building secure, full-stack AI applications using Amplify and Amazon Bedrock. If you want to jump right in, you can check out our getting started guide or sample buybacks.
Add AI capabilities
To create a full stack with AI capabilities in Amplify AI kit, you'll need to define AI "routes" in Amplify Data Schema, along with your data model and custom queries. AI routing is like an API endpoint for interacting with back-end AI functions. There are currently two paths to AI:
Conversation: Conversation Routing is a real-time, multi-turn API. Conversations and messages are automatically stored in DynamoDB, and responses are transmitted to the client in real-time. An example of this is any chat-based AI experience or conversational user interface.
Generation: A simple request-response API. The Generate route is an AWS AppSync query that generates structured data based on your definition. Common uses include generating structured data from unstructured inputs and summaries.
Serverless full stack with AI architecture
The scaling full stack with AI is completely serverless. It's fast, and with Amplify sandboxes, you can iterate on the back-end code and see updates in a real cloud environment. With Amplify Sandbox, you get an on-premises development experience for cloud resources. Updating the back end and front end as you develop features means you can build functionality end-to-end without developing silos. When you're ready to deploy, push a branch to scale up the host, which will deploy your back end and front end at the same time.
AWS AppSync API: full stack with AI gateway for secure real-time connection to clients. The AppSync API acts as the hub for everything: data requests from clients or LLMS go through AppSync.
Lambda function: acts as a bridge between AppSync and Amazon's bedrock. It retrieves the conversation history, calls Bedrock's streaming converse API, and processes the calling tool as an AppSync query.