IntelliJ Junie Configuration¶

This guide explains how to configure JetBrains Junie (the autonomous coding agent) to use the RIDDL MCP Server for AI-assisted domain modeling.

What is Junie?¶

Junie is JetBrains' autonomous AI coding agent that can understand your codebase, execute tasks, and make changes across multiple files. With the RIDDL MCP Server connected, Junie gains specialized capabilities for working with RIDDL domain models.

Prerequisites¶

IntelliJ IDEA with Junie support (MCP requires a recent version)
Junie enabled in your JetBrains account
riddlg installed and on your PATH (see Installation):
```
brew install ossuminc/tap/riddlg
```

Configuration¶

Junie launches MCP servers as local processes (stdio transport).

Settings UI Method¶

Open Settings (Cmd+, or Ctrl+Alt+S)
Navigate to Tools > Junie > MCP (or Model Context Protocol)
Click Add Server and configure:

Field	Value
Name	`riddl`
Command	`riddlg`
Arguments	`mcp`

JSON Method¶

Where Junie accepts a JSON server definition, use the standard stdio form:

{
  "mcpServers": {
    "riddl": {
      "command": "riddlg",
      "args": ["mcp"]
    }
  }
}

No URL and no API key are involved. If riddlg isn't on the system PATH, use its absolute path as the command.

Restart IntelliJ¶

After configuration:

Close IntelliJ IDEA
Relaunch IntelliJ IDEA
Open Junie to verify the RIDDL tools are available

Junie Capabilities with RIDDL¶

Autonomous Tasks¶

Junie can autonomously:

Validate all RIDDL files in your project
Fix validation errors across multiple files
Generate new entities, contexts, or domains
Refactor RIDDL structures
Add missing documentation

Example Autonomous Tasks¶

"Validate and fix my RIDDL model" — Junie scans all .riddl files, runs validation, fixes resolvable errors, and reports issues needing a human decision.

"Make this model simulation-ready" — Junie runs check-simulability, identifies missing handlers and state definitions, adds placeholders, and re-validates until the model is simulation-ready.

Usage Examples¶

Project-Wide Validation¶

Validate all RIDDL files in this project and create a report of any issues

Generate from Requirements¶

Create RIDDL definitions for a subscription management system with monthly and annual plans, automatic renewal, cancellation handling, and usage tracking

Refactoring¶

Move the Payment entity from the Orders context to a new Payments context, updating all references

Error Resolution¶

The validation shows 15 undefined reference errors. Fix them by adding the missing type definitions

Best Practices¶

Start small: Test Junie with simple RIDDL tasks first
Review diffs: Always review changes before applying
Use version control: Commit before running autonomous tasks
Provide context: Include relevant requirements in your prompts
Iterate: Break large tasks into smaller, verifiable steps

Troubleshooting¶

Junie Not Available¶

Verify Junie is enabled in your JetBrains account
Update to a version with MCP support

MCP Tools Not Working¶

Restart IntelliJ after configuration
Verify riddlg mcp starts from a terminal (it waits on stdin)
If riddlg isn't found, use its absolute path as the command
Check Junie logs: View > Tool Windows > Junie > Logs

Available RIDDL Tools¶

Tool	Use For
`riddl_validate`	Check RIDDL source for errors
`riddl_outline`	Summarize a model's definitions
`validate-partial`	Check incomplete models
`check-completeness`	Find missing elements
`check-simulability`	Verify simulation readiness
`map-domain-to-riddl`	Convert descriptions to RIDDL
`explain-error`	Understand validation errors
`suggest-next`	Get recommendations

See MCP Tools for the full catalog of 13.

Back to MCP Overview