About Diagram Bridge Mcp
Diagram Bridge MCP is an MCP server for LLMs that create diagrams. It combines format selection, format-specific guidance, and image rendering, supporting a workflow from choosing a format to producing an image.
Key features
- 01Recommend diagram formats from requirements
- 02Generate format-specific syntax guidance
- 03Render diagram code as PNG or SVG
- 04Support Mermaid, PlantUML, D2, and more
- 05Cache renders and save output files
Use cases
- 01Choose a format for a project architecture diagram
- 02Write diagram code using format-specific guidance
- 03Render diagram code to a PNG or SVG file
MCP support
The project exposes format recommendations, diagram-writing guidance, and rendering tools to MCP clients. MCP communication uses STDIO.
Configuration
Local server config is the same for every client. Copy it into your client’s MCP config file.
We haven’t collected a package or remote endpoint for this server. Follow the instructions in the project README.
Tools
Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.
Details
From the project READMEDiagram Bridge MCP Server
A comprehensive MCP server that bridges the gap between LLMs and diagram creation through three powerful, independent tools: intelligent format selection, format-specific instruction generation, and professional diagram rendering (using kroki server). Each tool can be used standalone or combined in sequence for complete diagram workflows - from choosing the right format to generating the final image.
Supported Formats: Mermaid, PlantUML, C4 Model, D2, GraphViz, BPMN, Structurizr, Excalidraw, Vega-Lite

Quick Start
-
Prerequisites:
- Node.js 18.0 or higher
- Docker and Docker Compose
-
Install and build:
git clone [email protected]:tohachan/diagram-bridge-mcp.git
cd diagram-bridge-mcp
npm install && npm run build
- Configure ports (optional):
# Copy and edit environment file to customize ports
cp .env.example .env
# Edit .env to set PORT and KROKI_PORT as needed
# Or set environment variables directly:
export PORT=4000 # MCP Server port (default: 3000)
export KROKI_PORT=9000 # Kroki service port (default: 8000)
- Start services:
# Option 1: Interactive port configuration
./scripts/start-with-ports.sh
# Option 2: Default ports (3000 for MCP, 8000 for Kroki)
docker-compose up --build
# Option 3: Custom ports via environment variables
PORT=4000 KROKI_PORT=9000 docker-compose up --build
- Add to Cursor (or any MCP client):
For direct local use (without Docker):
{
"mcpServers": {
"diagram-bridge-mcp": {
"command": "node",
"args": ["/path/to/diagram-bridge-mcp/dist/index.js"],
"enabled": true
}
}
}
For Docker-based setup (recommended):
{
"mcpServers": {
"diagram-bridge-mcp": {
"command": "docker",
"args": [
"exec",
"-i",
"diagram-bridge-mcp",
"node",
"dist/index.js"
],
"enabled": true
}
}
}
Note: The MCP protocol uses STDIO communication. The HTTP endpoints (:3000/health) are only for Docker monitoring, not for MCP client connections.
- Start using: The server provides three MCP tools ready for use!
💡 Usage Example
Here's a simple prompt to demonstrate the complete workflow:
"Choose the best diagram format for visualizing a current project architecture, then generate the diagram code and render it to an image file. Save the result as 'architecture-diagram'."
This single prompt will:
- Analyze your request and recommend the optimal format
- Get specific instructions for creating the diagram in that format
- Create and render the final diagram to a PNG/SVG file
🎨 Examples Gallery
Want to see what's possible? Check out Examples Gallery
🛠️ Tools Overview
help_choose_diagram
Purpose: Intelligent format selection based on your requirements
- Analyzes your request using AI-powered heuristics
- Provides confidence-scored recommendations
- Works with format constraints when needed
- Generates structured prompts for optimal LLM decision-making
Example: "I need to show database relationships" → Recommends Mermaid ER diagram format with detailed reasoning
get_diagram_instructions
Purpose: Generate format-specific coding instructions
- Creates tailored prompts for any supported diagram format
- Includes syntax guidelines, best practices, and common pitfalls
- Provides working examples and output specifications
- Ensures LLMs generate syntactically correct diagram code
Example: For Mermaid format → Detailed Mermaid syntax guide with examples
render_diagram
Purpose: Transform diagram code into professional images
- Renders diagrams via integrated Kroki service
- Supports PNG and SVG output formats
- Includes intelligent caching for performance
- Provides file storage with absolute paths for easy access
Example: Mermaid code → High-quality PNG/SVG image file
⚙️ Configuration
Basic Setup
- Default storage:
{project-root}/generated-diagrams/ - Default Kroki: Uses local Docker Kroki service
- Default ports: MCP Server (3000), Kroki (8000)
Port Configuration
The server automatically configures ports based on environment variables:
- External Port (
PORT): The port exposed by Docker container (default: 3000) - Kroki Port (
KROKI_PORT): The port for Kroki service (default: 8000) - Internal Communication: Services communicate internally using fixed container ports
# Interactive configuration (recommended)
./scripts/start-with-ports.sh
# Quick port change
PORT=4000 KROKI_PORT=9000 docker-compose up --build
# Check running services
curl http://localhost:4000/health # Using your custom PORT
Note: When running in Docker, the server provides HTTP endpoints for health monitoring at /health and service info at /info.
For advanced configuration, see our Configuration Guide.
📚 Documentation
- Quick Start: Port Configuration - Fast port setup guide
- Installation Guide - Detailed setup instructions
- Docker Deployment - Container deployment options
- API Reference - Complete tool and resource documentation
- Configuration Guide - Environment variables and MCP client setup
- Development Guide - Contributing and development setup
- Architecture - System architecture and design patterns
- Performance & Security - Performance metrics and security features
- Troubleshooting - Common issues and solutions
Perfect for: LLMs that need to create diagrams but don't know which format to use or how to render them professionally.
Powered by: Kroki - Universal diagram rendering service
Made with ❤️
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.