yenchieh/diagramzu-mcp
yenchieh/diagramzu-mcp is an open-source Model Context Protocol (MCP) server that bridges AI assistants and the Diagramzu diagramming platform. It enables AI models to programmatically create, read, update, and delete Mermaid diagrams directly within a shared workspace, eliminating the need to paste raw Mermaid syntax. The server exposes a set of tools that allow AI to generate flowcharts, sequence diagrams, entity-relationship diagrams, mind maps, and more, and then store them in a collaborative space where users can comment, embed, and present via clean, shareable URLs. Under the hood, the MCP server translates natural language instructions into Diagramzu API calls, managing authentication, workspace organization, and diagram lifecycle. Once generated, diagrams live on Diagramzu’s servers, providing a persistent, collaborative canvas for teams. AI can also read existing diagrams, appending comments or suggesting changes, enabling a conversational workflow around visual documentation. Key features include generating various Mermaid diagram types, reading back diagram contents for iterative editing, adding comments to diagrams for team discussion, embedding diagrams in external documents or websites via iframes or links, and presenting diagrams in a full-screen slide mode. The integration leverages MCP’s standard tool interface, making it compatible with any MCP-supporting AI client such as Claude Desktop, Continue.dev, or custom applications. Beyond basic diagramming, Diagramzu MCP supports versioning and workspace management, allowing multiple diagrams to coexist and reference each other. The open-source nature of the server means developers can self-host, extend with custom diagram types or integrations, and inspect the code. The project is actively maintained on GitHub, welcoming contributions and community feedback.
- Category
- Developer Tools
- Pricing
- Open Source
Problem solved
AI assistants typically generate Mermaid code that needs to be manually copied into an editor, rendered, and shared—a cumbersome process that breaks flow and makes collaboration difficult. Diagramzu MCP solves this by letting AI interact directly with a live diagramming workspace: it creates diagrams, organizes them in shared projects, and provides clean URLs for sharing, commenting, embedding, and presenting. This turns AI into a collaborative partner that can instantly produce visual documentation without interrupting the user's workflow.
Use cases
- An AI assistant generates a system architecture diagram during a design discussion and places it in a shared workspace for instant team review.
- A developer asks the AI to create a sequence diagram from a code snippet, then the AI reads it back, adds annotations, and shares a presentation link.
- During sprint planning, the AI produces a mind map of user stories in Diagramzu, and team members comment on each node to refine requirements.
- A technical writer uses the AI to draft an ER diagram for database documentation, then embeds it directly in the docs via a clean URL.
- An engineering manager asks the AI to create a flowchart of a deployment pipeline and present it in full-screen mode during a standup.
- A student learning data structures asks the AI to generate linked list operations as a flowchart and save it for later study, with the ability to comment on steps.
Audience
- Developers
- Product Managers
- Data Scientists
- Content Creators
Pros
- Seamless AI integration via the MCP standard, works with any MCP-compatible client.
- Eliminates manual copy-paste of Mermaid code, saving time and reducing friction.
- Collaborative workspace with commenting, embedding, and presenting on clean URLs.
- Supports a variety of diagram types: flowcharts, sequence diagrams, ER diagrams, mind maps.
- Open-source and self-hostable, allowing customization and security control.
Cons
- Requires a Diagramzu account and internet connectivity, creating an external dependency.
- Currently limited to Mermaid syntax, not extending to other diagramming languages like PlantUML or D2.
- Initial setup of the MCP server and API keys may be daunting for non-technical users.
Official website: https://github.com/yenchieh/diagramzu-mcp
Canonical LambdaBase page: https://www.lambdabase.com/tools/yenchieh-diagramzu-mcp