AI Assisted MDK: Faster Micro-Integration Development for Solace 

Solace is adding an AI-assisted workflow to the Micro-Integration Development Kit (MDK), letting developers — and autonomous AI agents — scaffold, implement, and test a production-ready Solace micro-integration through a single pipeline of Claude Code skills.

The Integration Problem Hasn’t Gotten Easier

The average enterprise now runs over 1,000 applications, roughly 90% of organizations operate across multiple clouds, “and the volume of data moving between systems is climbing about 25% year over year. Integration teams are being asked to connect more systems, faster, with less tolerance for the batch-oriented, tightly-coupled designs that used to be good enough.

Solace decomposes integration the same way modern architecture decomposed applications: instead of monolithic point-to-point connections, Micro-Integrations are small, purpose-built modules that each move data between one external system and the event mesh, with room for transformation, enrichment, or filtering along the way. It’s a pattern that scales — but every new Micro-Integration is still a new piece of software that has to be built, wired to the right binder, and tested — that’s the gap AI Assisted MDK closes.

Building event-driven integrations has typically meant a lot of upfront work before any real business logic gets written: boilerplate code, framework configuration, and manual project scaffolding. Solace’s AI Assisted MDK (Micro-integration Development Kit) is built to remove that overhead. It pairs Anthropic’s Claude Code with Solace’s MDK framework so developers — or fully autonomous AI agents — can go from zero to a production-ready Solace micro-integration in a fraction of the time.

Why it Matters

Scaffolding a new micro-integration by hand is repetitive work that doesn’t scale well as demand for new integrations grows. Every project needs the same boilerplate, the same test infrastructure, the same binder configuration — work that ties up engineering time without adding differentiated value. The AI Assisted MDK turns that fixed cost into a repeatable, automated step, which translates into a few concrete outcomes:

• Faster time to production: Projects that used to take days of setup can be scaffolded, implemented, and tested in a single pipeline run, so teams spend their time on integration-specific logic instead of project setup.
• Lower cost per integration: Because the skills produce consistent, tested output, teams need less senior engineering time per project — useful for organizations that need to stand up many similar integrations across teams or customers.
• More consistent quality: Every project comes out with the same Maven structure, the same TestContainers-backed test suite, and binders that follow the same patterns, reducing the variance you’d otherwise get across different developers or teams.
• Lower barrier to entry: Contractors, sales engineers, or developers newer to the MDK framework can produce production-ready output without first becoming experts in Spring Cloud Stream or the framework’s internal conventions.
• A path to unattended integration builds: Because the pipeline works the same way for a human developer or an autonomous agent, it opens the door to running integration builds as part of larger automated workflows, not just interactive development sessions.

How it Works

AI Assisted MDK is a set of structured AI “skills” — slash commands that each handle one phase of development and hand off to the next:

• /analyze-integration-tech — produces a technology overview for the integration
• /init-mi-project — scaffolds the base project
• /add-binder — configures the Spring Cloud Stream binder
• /createMi — generates the integration implementation
• /createMiTest — builds out the test suite

Each skill in the pipeline scaffolds a complete, tested, production-ready project. The tooling is built on Spring Cloud Stream and integrates with LangChain4J for LLM and vector store capabilities. Output includes Maven multi-module project structures, TestContainers-backed test suites, and Spring Cloud Stream binders refined through iterative AI passes — not just generated code, but code that matches the patterns the framework expects.

Who it’s For

AI Assisted MDK is built for developers, sales engineers, and anyone building on the MDK platform. The workflow is designed to produce the same consistent output whether a human developer is running the skills to speed up their own work or an AI agent is running the full pipeline unattended.

What’s Next

The current release covers the core scaffolding pipeline and soon the MDK will support Cloud Managed Micro-integrations, part of a broader push to bring AI-assisted tooling across the Solace developer experience.

Availability

The AI Assisted MDK is available now. Teams building event-driven integrations on Solace can start using it today to reduce the engineering time spent on scaffolding and testing, and redirect it toward the integration logic that’s specific to their business.

To learn more contact us.