Amazon Web Services is moving forward with its AI-powered Q services for developers, with the general availability of Q Developer, including the product formerly called CodeWhisperer, as well as a new service for building business apps now in preview.
CodeWhisperer, the FAQ states, is now part of Amazon Q Developer, with features including AI-powered code suggestions in the IDE or command line, security vulnerability scanning and remediation suggestions.
The Q Developer Agent is able to generate instructions, code and tests from natural language prompts, according to the documentation, working either in the Visual Studio Code IDE or in CodeCatalyst, an AWS DevOps service. There is also a code transformation feature, for actions such as upgrading a Java application to the latest language version.
CodeWhisperer was previously free for individual use, though the free tier was restricted to 50 security scans per month. That restriction remains in Q Developer Free Tier, which in addition has a limit of 50 interactions per month for chat, debug and tests, 5 agent tasks, 1,000 lines of code transformation, and 25 queries about AWS account resources. In this respect, though, it is ahead of GitHub’s Copilot which has no free tier.
The Pro version at $19 per user/month has unlimited chat and tests, and higher limits on other services, with 30 agent tasks, 4000 lines of code transformation, and 500 security scans. There is an additional service for generating SQL queries.
Another nuance is that Q Developer pro offers indemnity for its output, in case of inadvertent use of licensed code, which the free tier lacks.
The documentation states that “all the features of CodeWhisperer are moving to Amazon Q Developer” but this does not apply to code customization “Amazon Q Developer will not provide customized code recommendations based on your internal code bases, even if you’re using customizations through CodeWhisperer,” the FAQ states, the justification perhaps being that this was a feature in preview.
New in preview is Amazon Q Apps, part of Amazon Q Business. Q Business, which aside from Apps is now generally available is an AI chat service informed by internal business data, with features such as summarizing reports or answering questions. Q Business can also perform actions such as drafting emails or, via plugins, submit leave requests. There are connectors to data sources including S3 storage, Microsoft SharePoint and Salesforce.
The new Apps feature generates applications from descriptions in natural language, or based on Q Business chats. Judging by the examples, the essence of Apps is that it can generalise the prompts used for a specific case – an example being to create an onboarding plan for a new employee – so that it becomes an app where the user enters key details such as name and role into a form.
Technically, Q Apps are made up of a collection of cards, where a card can take user input and generate text or actions. Available actions are determined by plugins. There are built-in plugins for Jira, Salesforce, ServiceNow and Zendesk allowing Q to create new issues, or organizations can create custom plugins based on an OpenAPI schema.