Step 1: Install the Extension Pack#

Open VS Code, go to Extensions (Ctrl+Shift+X), and search for “Anypoint Code Builder” or “MuleSoft”. Install the official extension pack from Salesforce:

1
# Or install via command line:
2
code --install-extension salesforce.mule-dx-pack

The pack includes:

• mule-dx-pack — Core Mule development (XML editors, canvas, debugger)
• mule-dx-data-weave-client — DataWeave language support with preview, debug, and test
• mule-dx-api-designer — RAML and OAS API specification editor
• mule-dx-deploy — One-click deployment to CloudHub, RTF, or on-prem
• mule-dx-vibes — MuleSoft Vibes AI assistant (optional)

Step 2: Configure Java and Mule Runtime#

After installation, ACB will prompt you to configure:

1
1. Open Command Palette (Ctrl+Shift+P)
2
2. Run: "Mule: Set Java Home"
3
3. Point to your JDK 17+ installation (e.g., /usr/lib/jvm/java-17-openjdk)
4
4. Run: "Mule: Set Runtime Home"
5
5. Point to Mule 4.6+ runtime directory

Verify with:

• Open any .xml file — syntax highlighting and validation should activate
• Status bar shows Mule runtime version

Step 3: Login to Anypoint Platform#

1
# Command Palette → "Mule: Login to Anypoint Platform"
2
# Or click the user icon in the activity bar

This connects ACB to:

• Anypoint Exchange (shared assets, connectors, templates)
• Anypoint Design Center (API specs)
• Anypoint Monitoring (for deployed apps)
• Anypoint Visualizer (for agent interactions)

Using a Template (Recommended)#

1
1. Command Palette → "Mule: Create Mule Project"
2
2. Select template: "HTTP API with Database" or blank project
3
3. Name: "hello-mule-integration"
4
4. Choose output directory

ACB generates a complete project structure:

1
hello-mule-integration/
2
├── src/
3
│   ├── main/
4
│   │   ├── mule/
5
│   │   │   └── hello-integration.xml    → Main flow definition
6
│   │   └── java/                        → Custom Java classes (optional)
7
│   └── test/
8
│       └── munit/
9
│           └── hello-integration-test.xml → MUnit test
10
├── pom.xml                              → Maven project config
11
├── exchange.json                        → Exchange metadata
12
└── .vscode/
13
    ├── settings.json                    → ACB project settings
14
    └── launch.json                      → Debug configuration

Visual Canvas vs. Code View#

ACB gives you two ways to build flows:

Canvas View (flow.xml → Design tab):

• Drag-and-drop connectors, transformers, and routers
• Visually wire flows with the same palette you know from Studio
• Click any component to configure it directly on the canvas
• Dynamic context menus filtered by project type

Code View (flow.xml → Source tab):

• Raw XML with language server validation
• Autocomplete for all Mule components
• Real-time error highlighting
• Code snippets for common patterns

Toggle between them with one click. Changes in either view are reflected in the other — no sync issues.

Building a Simple HTTP → Database Flow#

Here’s a practical example — accept HTTP POST, transform JSON, insert into database:

1. Canvas Design: Drag these components from the palette:

• HTTP Listener (config: port 8081, path /create)
• Set Payload (DataWeave transformation)
• Database Insert (config: your DB connection)

2. DataWeave Transformation:

1
%dw 2.0
2
output application/json
3
---
4
{
5
  "fullName": payload.firstName ++ " " ++ payload.lastName,
6
  "email": payload.email as String,
7
  "createdAt": now() as String {format: "yyyy-MM-dd'T'HH:mm:ssZ"},
8
  "source": "ACB-via-HTTP"
9
}

3. Connect to Database:

• Add a Database connector from the palette
• Configure connection (MySQL/PostgreSQL/Oracle)
• Set SQL: INSERT INTO users(name, email, created_at, source) VALUES(:name, :email, :createdAt, :source)
• Map DataWeave output fields to SQL parameters

Running Locally#

1
1. Open the Run and Debug panel (Ctrl+Shift+D)
2
2. Select "Launch Mule Application" configuration
3
3. Press F5 or click the green play button
4
4. ACB starts Mule runtime, deploys your app, and shows logs in the Debug Console
5
5. Test: curl -X POST http://localhost:8081/create -H "Content-Type: application/json" -d '{"firstName":"John","lastName":"Doe","email":"john@example.com"}'

The debugger supports breakpoints in both XML and DataWeave editors. Set breakpoints, inspect payload variables, and step through flows.

Inline DataWeave Preview#

When editing DataWeave expressions, ACB shows a live preview panel:

1
%dw 2.0
2
input payload application/json
3
output application/json
4
---
5
{
6
  "user": payload.firstName ++ " " ++ payload.lastName,
7
  "isActive": payload.status == "active",
8
  "tags": if (payload.role == "admin") ["admin", "power-user"] else ["standard"]
9
}

• Input tab: Paste a sample JSON/XML/CSV payload
• Output tab: See the transformation result update in real-time
• Warnings: ACB highlights type mismatches, null-safety issues
• Performance hints: Shows estimated complexity for large payloads

Graphical Data Mapping (New in Feb 2026)#

The February 2026 release introduced graphical DataWeave mapping — drag source fields to target fields, and ACB generates the DataWeave expression automatically:

1
Source (JSON):                      Target (CSV):
2
┌─────────────────────┐            ┌─────────────────────┐
3
│ firstName ──────────┼───────────▶│ name                 │
4
│ lastName  ──────────┼───++──┘    │                      │
5
│ email     ──────────┼───────────▶│ email                │
6
│ role      ──────────┼───────────▶│ role                 │
7
│ status    ──────────┼───if──┐   │ active               │
8
└─────────────────────┘        │   └─────────────────────┘
9
                               │   ┌─────────────────────┐
10
                               └──▶│ if: status == "active" then "Yes" else "No"
11
                                   └─────────────────────┘

For complex transformations, the graphical view is a massive time-saver. For simple ones, inline DW is faster. ACB supports both.

DataWeave Testing#

ACB includes a built-in DataWeave test runner:

1
%dw 2.0
2
import * from dw::test
3
---
4
describe "User Transformation" do {
5
  it "concatenates first and last name" do {
6
    input = {firstName: "John", lastName: "Doe"}
7
    expected = "John Doe"
8
    actual = do {
9
      input.firstName ++ " " ++ input.lastName
10
    }
11
    expect(actual, expected)
12
  }
13

14
  it "handles null gracefully" do {
15
    input = {firstName: null, lastName: "Doe"}
16
    expected = " Doe"
17
    actual = do {
18
      (input.firstName default "") ++ " " ++ (input.lastName default "")
19
    }
20
    expect(actual, expected)
21
  }
22
}

Run via: Right-click .dwl file → “Run DataWeave Test”

Setting Breakpoints#

• XML flows: Click the line number gutter in flow XML
• DataWeave: Click line numbers in DW editor
• Conditional breakpoints: Right-click breakpoint → “Edit Breakpoint” → set condition

Debug Session#

When you hit a breakpoint:

1. Variables panel shows the current Mule message — payload, attributes, variables
2. Call stack shows flow execution path with XML line numbers
3. Watch expressions let you evaluate DataWeave expressions on-the-fly
4. Step controls: Step Over (F10), Step Into (F11), Continue (F5)
5. Inline values — ACB shows variable values directly in the XML editor next to the component

Debug Console#

The Debug Console shows:

• Mule runtime logs with color-coded levels
• DataWeave evaluation results
• HTTP request/response tracing
• Real-time error stack traces with clickable source links

Option 1: Deploy to CloudHub (Easiest)#

1
1. Command Palette → "Mule: Deploy to CloudHub"
2
2. Select target environment (DEV/QA/PROD)
3
3. Configure: app name, region, vCore size, workers
4
4. Click Deploy

ACB handles: packaging → uploading → deploying → health check. Took me 2 minutes on a test run.

Option 2: Deploy to Runtime Fabric#

1
1. Command Palette → "Mule: Deploy to Runtime Fabric"
2
2. Select RTF target (Kubernetes-based)
3
3. Configure: replicas, CPU/memory limits, ingress
4
4. Optional: set environment variables for config injection

Option 3: Export for CI/CD#

1
1. Command Palette → "Mule: Export Deployable Archive"
2
2. Generates: target/hello-mule-integration-1.0.0-mule-application.jar
3
3. Use in your pipeline: mvn clean package -DskipTests

ACB generates standard Maven projects — compatible with any CI/CD pipeline (GitHub Actions, Jenkins, GitLab CI).

Unified Workspaces#

• Import existing Anypoint Studio workspaces directly — ACB creates pointers to your files (no duplication)
• Multi-root workspace support: DataWeave and API components work across multiple projects in one workspace
• Context-aware command palette: shows only relevant actions for the current file/project

Modernized Canvas#

• Updated visual design with improved hierarchy and context awareness
• Less rigid, more fluid feel — closer to modern dev tools
• Configuration tabs accessible directly from canvas components
• Reduced visual clutter, improved focus on flow logic

Custom Metadata Editor#

• Define input/output schemas visually without leaving the canvas
• Graphical custom metadata editor for complex data structures
• Real-time validation against Exchange-defined types

Graphical DataWeave Mapping (mentioned above)#

• Drag-and-drop field mapping with automatic DW expression generation
• Works for JSON, XML, CSV, and flat file transformations

Simplified Error Handling#

• Errors consolidated into unified view (no more hunting across panels)
• Faster scan and action on runtime errors during development
• Clearer error categorization with fix suggestions

AI: MuleSoft Vibes#

ACB includes MuleSoft Vibes — an AI assistant trained on MuleSoft documentation, common patterns, and the Exchange ecosystem:

• Natural language to flow: “Create an HTTP endpoint that accepts orders, validates them, and inserts into Salesforce” → generates the full XML flow
• DataWeave generation: “Transform this Salesforce Account to a CSV row” → generates DataWeave expression
• Error resolution: Paste an error → Vibes suggests fixes with code examples
• Documentation lookup: “How do I configure OAuth 2.0 for Salesforce connector?” → answers with exact config steps