Development Agent

The Development Agent is an AI-driven assistant for designing and scaffolding Axon Framework applications inside Axoniq Platform. You start with a prompt describing what you want to build, and the agent walks you through requirements, user journeys, components, and finally generates working code, keeping you in control at every step.

From prompt to code

The agent works in stages. Each stage produces an artifact that you review and refine before moving to the next:

  1. Requirements: the agent turns your initial prompt into a structured list of requirements. You can edit, add, or remove individual requirements.

  2. User journeys: the agent derives end-to-end journeys from the requirements, including primary and alternative flows. Journeys describe the behavior the application should exhibit.

  3. Components: the agent breaks the journeys down into Axon Framework components: aggregates, command handlers, event handlers, query handlers, and projections. You can ask the agent to rename, simplify, merge, or split components.

  4. Code: the agent generates a working project. You can download the project as a zip and run it locally, or continue iterating inside the platform.

At any point you can use the Magic Wand control next to a journey, step, component, or other artifact to ask the agent for a targeted change ("simplify this", "rename to a clearer term", "remove this"). Suggestions are sent back into the conversation and applied incrementally.

Code generation

The Development Agent generates idiomatic Axon Framework code. Both Java and Kotlin are supported as target languages, and you can pick the language when starting a project. The generated code follows Axon Framework conventions for command handlers, event sourcing, and query handlers, and ships with a Spring Boot entry point and the necessary configuration to connect to Axoniq Platform.

The agent runs scheduled tasks for long-running steps such as code generation and bulk modifications. You can follow progress in the Tasks panel of your project. Failed tasks can be retried, and root operators can trigger a fresh code generation if a project ends up in a bad state.

Sharing projects

By default, projects are private to your workspace. You can make a project public to share it through the marketplace, or invite specific collaborators. See Sharing projects for details.

Limits

Each workspace has a number of AI credits per billing period, which are consumed when the agent works on your behalf (analyzing prompts, generating requirements, generating code, etc.). The amount of available credits and your current usage are visible in the workspace switcher and on the Subscription page. Higher-tier plans include more credits, and additional credits can be purchased as add-ons. See Subscriptions for plan details.

The number of private projects per workspace is also bounded by your subscription. Public projects don’t count toward this limit.