Support for Inline Code Nodes in Amazon Bedrock Flows in Public Preview

Amazon has today unveiled the public preview of online code nodes in Amazon Bedrock Flows, a feature designed to make it easier for developers to integrate Python scripts directly into their workflows. This innovation promises to simplify the development of generative artificial intelligence applications by eliminating the need to create separate Lambda functions for simple logical tasks, optimizing preprocessing and postprocessing tasks like data normalization and response formatting.

Organizations using Amazon Bedrock Flows will now have the capability to design and deploy more scalable and efficient workflows in the generative AI space. Some of the main benefits include transforming input data without the need for additional Lambda functions, allowing, for instance, the extraction of specific fields from JSON and preparing data for large language models (LLMs). Additionally, users will be able to perform operations directly on model outputs, such as extracting entities from responses and formatting JSON for downstream systems.

With this functionality, users of Amazon Bedrock Flows will be able to manage complex AI use cases that require multiple stages, integrating popular packages like OpenCV and SciPy. The experience has been streamlined, enabling developers to create and manage inline code through the AWS console or the Amazon Bedrock API. Complete traceability will also be included, making it easier to monitor inputs and outputs.

A notable example of this application is Thomson Reuters, which has managed to optimize flow management, allowing over 16,000 users to efficiently handle thousands of workflows without needing to create Lambda functions for each block of code. This improvement promotes flexibility in data processing and the creation of complex flows through a self-service interface.

The new functionality is available in various AWS regions, including US East (Northern Virginia, Ohio), US West (Oregon), and Europe (Frankfurt). Developers can access the Amazon Bedrock console or APIs to start building workflows with these new capabilities. With this enhancement, it is expected that more companies will adopt generative AI solutions, accelerating iteration cycles and increasing participation in application development in this field.