The Mendix Context Bridge server enables AI agents to inspect and understand Mendix low-code application structures by parsing .mpr project files in both offline and cloud-based modes.
Core Capabilities:
• List modules - Get an overview of all modules in the opened .mpr file
• Inspect microflows - Find microflows by name and retrieve their logical steps and execution flow
• Extract domain models - View all entities and types within specific modules
• Examine database schema - Inspect the internal SQLite database structure to understand how Mendix stores project metadata
• Discover project components - Automatically detect projects and list documents (Microflows, Pages) with optional module filtering
Operating Modes:
• Offline mode - Fast local parsing using the "Shadow SDK" for direct binary .mpr and .mxunit file analysis without requiring cloud connectivity
• Cloud mode - Precise inspection via the official Mendix Platform SDK for 100% accurate representations
AI Integration:
• AI-ready output - Transforms complex Mendix objects into clean, circular-safe JSON that AI agents can easily process • Privacy-first - Operates entirely offline to keep intellectual property secure on your machine • Low-code bridge - Makes proprietary Mendix file formats accessible to AI tools like Claude, Cursor, and other MCP-compatible clients
Connects to Mendix project .mpr files (which use SQLite format) in read-only mode to query project structure, entities, microflows, and modules, enabling AI agents to understand and analyze local Mendix application models.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Mendix Context Bridgeshow me all microflows in the OrderManagement module"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Mendix MCP Server
A powerful bridge between Mendix Applications and AI Agents.
The Mendix Local MCP Server is a Model Context Protocol (MCP) server designed to empower AI tools (like Google Antigravity, Claude Desktop, or Cursor) to inspect, read, and understand the structure of Mendix projects.
It operates in a unique Dual Mode:
Shadow SDK (Local): Instant, offline access to
.mprfiles via direct binary parsing.Official SDK (Cloud): Deep, accurate inspection using the Mendix Platform SDK.
Why Use This Tool?
Integrating Low-Code platforms with AI agents is notoriously difficult due to proprietary file formats. This tool solves that problem.
🚀 AI-Ready: Exposes complex Mendix logic (Microflows, Domain Models) as clean, AI-readable JSON.
⚡ Blazing Fast (Local): Uses a custom "Shadow SDK" to parse
.mprSQLite databases and.mxunitbinaries without waiting for the Model SDK to load.🛡️ Circular-Safe: Automatically handles the notorious "Circular Structure" errors common in the Mendix SDK by applying a DTO Transformation Layer.
🔒 Privacy-First: Can operate entirely offline (Local Mode), keeping your intellectual property on your machine.
Key Features
Project Discovery: Automatically detects Mendix projects in your workspace.
Module Browsing: Recursively filters documents (Microflows, Pages) by Module.
Shadow Parsing: Extracts metadata from binary blobs without the overhead of the full SDK.
Official SDK Integration: Fetches authoritative data from the Mendix Cloud when absolute precision is required.
DTO Sanitization: Maps complex Mendix objects to flat, safe JSON for AI consumption.
Installation
Prerequisites
Mendix Studio Pro v10.24.13 or newer (Required for the new SQLite-based
.mprformat)Node.js (v18 or higher)
A local Mendix project (Git-backed or local file)
Setup
Clone the repository:
git clone https://github.com/YourUsername/mendix-local-mcp.git cd mendix-local-mcpInstall Dependencies:
npm installBuild the Project:
npm run build
Configuration
To unlock the full power of the Official SDK (Cloud Mode), you must configure your Mendix Personal Access Token (PAT).
Add the server to your MCP Client configuration (e.g., mcp_config.json for Antigravity or Claude Desktop):
Note: Generate your PAT in the Mendix Developer Portal with the scope
mx:modelrepository:repo:read.
Usage
Once running, the server exposes the following tools to your AI Agent:
Local Mode Tools (Offline)
These tools use the "Shadow SDK" and do not require a token.
list_local_documents(module_name?)Lists all documents in the project. Optional filter by module.
get_domain_model(module_name)Extracts a simplified Domain Model using parsing.
inspect_local_microflow(microflow_name)Reads binary definitions to show microflow logic.
inspect_database_schema(table_name?)(Debug) Inspects the internal
.mprSQLite schema.
Cloud Mode Tools (Online)
These tools use the Official Mendix SDK.
get_module_entities_sdk(module_name, project_id, branch?)Recommended for Refactoring. Fetches a 100% accurate, sanitized JSON representation of the Domain Model from the Mendix Cloud.
Project Structure
Contributing
Contributions are welcome! If you'd like to improve the Binary Parser or add more DTO mappers:
Fork the repository.
Create a feature branch (
git checkout -b feature/amazing-feature).Commit your changes (
git commit -m 'Add some amazing feature').Push to the branch (
git push origin feature/amazing-feature).Open a Pull Request.
License
This project is licensed under the MIT License - see the LICENSE file for details.