This release tackles three things that slow down Mule developers every day: a canvas that requires click-throughs to understand, single-project-only local runs, and a debugger that got in the way more than it helped.
Descriptive UI Layout, Multi-App Runtime, and Debugger Enhancements
Descriptive UI layout gives every component a plain-English description so you can read a flow like a recipe. Multi-App Run Configurations let you deploy multiple projects to a single local runtime with an AI skill that configures it conversationally. And a set of debugger fixes eliminate wrong-file navigation, unreadable payloads, and missing execution tracking on the canvas.
Descriptive UI layout: Read your flows like a recipe
The traditional Mule flow canvas tells you what components are in a flow, but not what they do. You end up clicking into each one to understand the logic – especially when every other node just says “Transform Message.” Descriptive UI Layout replaces the node-graph mental model with a vertical, recipe-style canvas. Each component renders as a card with an AI-generated, human-readable description of what it actually does in business terms.
A Transform Message processor no longer just says Transform Message; it says: “Stores core opportunity fields in a variable for later NetSuite processing.” A Flow Reference doesn’t just show the target flow name; it explains what happens when that flow runs.
How it works
MuleSoft Vibes uses the component’s type, operation, configuration values, and surrounding data flow context to generate descriptions. For connectors, it reads the operation and object type. For DataWeave transforms, it inspects input/output schemas. For Choice routers, it translates conditions into plain English business rules.
The layout itself follows a top-to-bottom vertical flow. Linear sequences render as a step list. Branching (Choice, Scatter-Gather) shows as indented sub-recipes under the parent step. Error handlers appear inline with a distinct background. Sub-flows are collapsible.
The trade off: Descriptive UI Layout optimizes for readability over density. If you have a flow with 30+ processors and prefer seeing the entire topology at a glance, the classic canvas layout remains available. Descriptive UI is best suited for flows where understanding what’s happening matters more than seeing everything at once.
Multi-app Run Configurations: Locally run your API-led stack
Mule projects don’t run in isolation. APIs call databases, events chain across apps and agents invoke MCP tools. But until now, testing these interactions locally meant juggling separate runtimes, separate terminals, and manual coordination across all of them. Take API-led connectivity as a common example: your experience-api, process-api, and system-api all need to be running simultaneously to validate an end-to-end request. That’s three runtimes, three sets of logs, and a lot of context switching.
This release introduces Run Configurations in ACB: reusable, named configurations that define which projects to deploy together, which Mule Runtime and JDK version to use, and what program/VM arguments to pass.
What a Run Configuration defines:
- Projects to run: select multiple Mule projects from your workspace
- Debug or Run Mode
- Mule Runtime version: Dropdown of locally installed runtimes (eg 4.6.0)
- Java version: Aligned with project requirements
- Program arguments and VM arguments: Separate fields with defaults pulled from your existing settings
Execution model: When you run a multi-project configuration, ACB packages each project via Maven, copies the JARs to the runtime’s apps/ folder, and starts a single shared runtime instance that deploys all applications on startup. One runtime, multiple apps, one click.
The Vibes skill: Rather than clicking through a UI form, you can create and manage Run Configurations conversationally through the Dev Agent. Say “Create a debug config for mule-ref-db-sapi and customer-api on runtime 4.6.0” and the agent generates the configuration, stores it in your .code-workspace file (multi-root) or launch.json (single project), and offers to run it immediately.
Configurations are also available through the VS Code Run and Debug dropdown and the Explorer context menu under the “Mule” section.
Debugger Enhancements for reliable troubleshooting
The ACB debugger has the architecture right: DAP protocol, transparent proxy, standard VS Code integration. This release enhances data display, navigation, and canvas integration
Data display
The Variables panel now shows you what you’d expect to see, not what the JVM happens to serialize. JSON payloads copy as valid JSON you can paste directly into Postman without manually stripping escape sequences. Non-Java MIME types (application/json, text/xml, text/plain) render as readable text instead of raw byte arrays.
When paused inside an error handler, a dedicated error scope surfaces the exception type, message, and cause at the top level – no more drilling through nested Java nodes to understand what failed. Variable values are editable mid-session (double-click, type, resume), and exception messages on error breakpoints are now immediately legible without expansion. It only evaluates payloads and variables, and can also evaluate DataWeave expressions. This is a critical capability that developers could leverage while debugging their expressions written in DataWeave.
navigation
Multi-config project debugging now works correctly. Hitting a breakpoint in config-b.xml opens config-b.xml – not whichever file happened to be active. Stepping into a
Canvas integration
The canvas now participates in the debug session. When execution pauses, the active processor gains a distinct dotted-outline highlight. You can see where you are without cross-referencing line numbers. If the paused component is off-screen, the viewport automatically pans to bring it into view. Both indicators are clear on resume or session termination. No more hunting through a large flow to find where execution stopped.
Get started with the latest ACB updates
Download your latest ACB Extension pack from Visual Studio marketplace today. So, check out the demo video to get a closer look at everything discussed above:
