Local-First Generation

RIDDL Generator

riddlg is a single binary that runs entirely on your machine. From a RIDDL model it generates documentation, API specifications, and code; from a brief it generates validated RIDDL — and it doubles as an MCP server for Claude and other AI assistants. No cloud service, no latency.

Download riddlg Documentation

What riddlg Generates

Point it at a RIDDL model and pick a target. Documentation, API specs, and AI RIDDL generation are free; code generation is a Pro feature.

Documentation (Free)

AsciiDoc and MkDocs sites with Mermaid diagrams, generated straight from your model.

API Specifications (Free)

OpenAPI, gRPC, and Smithy specifications generated from your contexts.

AI RIDDL Generation (Free)

Turn a natural-language brief into validated RIDDL using a local AI model.

Code Generation (Pro)

Runnable Quarkus source with AI-filled, compile-verified handler bodies.

Also an MCP Server for Your AI Assistant

The Model Context Protocol (MCP) is an open standard that lets AI assistants access external tools and data. Run riddlg mcp and riddlg becomes a local MCP server, giving assistants like Claude direct access to RIDDL's validation, analysis, and generation capabilities.

This used to be a hosted service; now it's just a subcommand of the binary you already have. You can hold natural-language conversations about your domain model while the AI validates syntax, checks completeness, suggests improvements, and generates code — all against a server running on your own machine.

13 Specialized Tools

In mcp mode, riddlg exposes tools that give AI assistants deep understanding of RIDDL syntax, semantics, and best practices.

Validation

  • validate-text
  • validate-url
  • validate-partial

Analysis

  • check-completeness
  • check-simulability
  • map-domain-to-riddl
  • explain-error

Guidance

  • suggest-next
  • expand-pattern

Template

  • list-patterns
  • list-templates
  • get-template
  • generate-test-cases

Built-in Knowledge

riddlg ships with comprehensive RIDDL documentation as MCP resources that AI assistants can access. This means the AI understands RIDDL deeply — not just syntax, but patterns, idioms, and best practices.

When you ask the AI for help, it draws on this knowledge to provide accurate, contextual assistance specific to RIDDL.

The result: richer, more accurate, more complete models so your design intent transmits faithfully to human audiences and AI implementers alike.

Available Resources

  • EBNF Grammar
  • Language Guide
  • Cheat Sheet
  • Patterns Catalog
  • Idioms Reference
  • Anti-Patterns Guide
  • Example Models

What Can You Do?

💬

Natural Language Modeling

Describe your domain in plain English. riddlg translates it to valid RIDDL.

✅

Real-Time Validation

Get instant feedback on syntax errors, missing references, and semantic issues.

📚

Docs & API Specs

Generate documentation sites and OpenAPI/gRPC/Smithy specs from your model.

🧩

Patterns & Idioms

Apply proven DDD patterns and RIDDL idioms to your domain with AI guidance.

Quick MCP Setup

1

Install riddlg (Homebrew, or a native/JVM download) so riddlg is on your PATH

2

Add riddlg to your Claude configuration file:

Claude Desktop / Claude Code Configuration

{
  "mcpServers": {
    "riddl": {
      "command": "riddlg",
      "args": ["mcp"]
    }
  }
}
3

Restart Claude and start asking about your domain model

Using a different AI client? The same riddlg mcp stdio server works with Cursor, Windsurf, and other MCP-compatible tools — see the riddlg documentation .

Prefer a Visual Editor?

Synapify wraps the same generation and AI capabilities in a desktop app, with a visual editor and simulation. Or stay on the command line with riddlg — both are free to start.

Code generation and Synapify simulation are Pro features ($29/month); everything else is free.

Download riddlg View Pricing