OpenFlowKit MCP Server

Community

Local-first open-source diagramming tools for Claude, Cursor, Windsurf & other MCP clients.

How to install ↓ Compare GitHub source repository ↗

Category: Dev Tools #26 of 438
Stars: ★ 848 Very popular
Transport: stdio (local process)
Runtime: Node.js
Credentials: No credential needed
License: MIT
Last commit: 8 days ago
Tools: 6

67FMRS · C

OpenFlowKit MCP server is a lightweight local extension of the diagramming tool, offering several useful utilities for integrating diagram creation into AI workflows. Installation is straightforward, but functionality is limited to the current version; some advanced features are not yet implemented.

Strongest · Maintenance 16/20 Weakest · Reliability 7/20

Reliability

7/20

Security and permissions

16/20

Maintenance

16/20

Documentation

15/20

Setup experience

13/20

Why each score

Reliability 7/20: Evidence: Repository has a quality CI workflow (quality.yml), unit tests (Vitest), e2e tests (Playwright), and MCP server source (mcp-server/) is present and npm-published. However, the server manifest lacks full tool/resource/prompt listings, and tests do not explicitly cover MCP server key paths. Deductions: No execution verification, static review caps at 12; insufficient test coverage and error-handling documentation.
Security and permissions 16/20: Evidence: README claims local-first, no server-side storage, API keys remain in browser and sent directly to providers; MCP server tools are deterministic local validation, templates, icon lookup, no external network or remote side effects. No malware, token leaks, or dangerous operations found. Deductions: Permission boundaries and confirmation mechanisms not explicitly documented; data exfiltration details not verified, but strong local execution and least-privilege design justify high score.
Maintenance 16/20: Evidence: Active project with 673 stars, 4 open issues, MIT license, CI present, recent releases (MCP server v0.1.2). README includes contributing guide and security policy. Deductions: Release history and commit frequency are inferred statically, no dates verified; dependency update policy not documented.
Documentation 15/20: Evidence: README extensively describes MCP server features (8 local tools, 5 resources, 3 prompts) and provides client config example (npx). A separate MCP server README is implied. Deductions: Missing detailed per-tool parameter reference; troubleshooting/error-handling sections incomplete; tool table not confirmed from files.
Setup experience 13/20: Evidence: MCP server is npm-published (@vrun-design/openflowkit-mcp), configuration is simple via npx; README provides JSON config for Claude/Cursor/Windsurf. Root project setup is straightforward (npm install && npm run dev). Deductions: No dedicated installation doc, no troubleshooting guide; platform compatibility unclear; not executed due to static review, so not full marks.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local files

Best for

Developers wanting to integrate AI-assisted diagramming into their workflow.
Teams that prioritize data privacy and prefer local tools.
Workflows that benefit from tight coupling between code and diagrams.

Not for

Users needing real-time multi-user collaboration (collaboration is experimental and off by default).
Users wanting GIF export (currently only MP4 is supported).
Users needing a desktop application (not yet released).

Required permissions

No API key required.
No network requests needed; all operations run locally.
May access local file system for codebase analysis.

Risks and side effects

Early-stage software; may have bugs or changes in functionality.
Codebase analysis may be inaccurate; AI-generated diagrams may require manual review.
Local-first means data stays in the browser local storage; sharing requires exports or viewer links.

Setup

Before you start

Runtime：Node.js

Add the following configuration to your MCP client:

{
  "mcpServers": {
    "openflowkit": {
      "command": "npx",
      "args": ["-y", "@vrun-design/openflowkit-mcp"]
    }
  }
}

Then restart the client to use the tools.

claude_desktop_config.json

{
  "mcpServers": {
    "openflowkit": {
      "command": "npx",
      "args": ["-y", "@vrun-design/openflowkit-mcp"]
    }
  }
}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "openflowkit": {
      "command": "npx",
      "args": [
        "-y",
        "@vrun-design/openflowkit-mcp"
      ]
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add openflowkit -- npx -y @vrun-design/openflowkit-mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After adding the npx -y @vrun-design/openflowkit-mcp config to Claude Desktop, Cursor, or Windsurf, the tool list should show local tools such as validate_dsl, create_viewer_url, analyze_codebase, find_icon_slug, fetch_template, and inspect_capabilities; calling inspect_capabilities confirms the connection.

Troubleshooting

If the MCP server fails to connect, check:
1. Ensure Node.js is installed and npx is available.
2. Check network connectivity to allow npx to download the package.
3. Reload the MCP client configuration to ensure it is correct.

Things to try

Once connected, you can ask your AI assistant things like:

Read the OpenFlowKit DSL cheatsheet, create a checkout flow with a promo-code branch and a Stripe webhook step, validate it, and create a viewer URL.
Analyze my codebase and generate an architecture diagram for it.
Find the icon slug for PostgreSQL and Redis nodes.
Fetch a flowchart starter template and validate my DSL.

Tools 6

validate_dsl read-only

Validates OpenFlowKit DSL syntax to ensure diagram descriptions are correct.

create_viewer_url read-only

Generates a viewer URL to open and edit diagrams in the browser.

analyze_codebase read-only

Analyzes a codebase to extract architecture information and generate diagrams.

find_icon_slug read-only

Searches the icon catalog for matching icon slugs for automatic icon assignment.

fetch_template read-only

Fetches available diagram templates as starting points for creations.

inspect_capabilities read-only

Inspects the MCP server's capabilities, listing available tools and resources.

Use cases

Create system architecture diagrams or flowcharts via natural language prompts in Claude or Cursor.

Analyze an existing codebase to produce architecture diagrams for better understanding.

Use templates to quickly generate standard diagrams, improving documentation efficiency.

Supported clients

Claude Desktop

Cursor

Windsurf

Listed from the project's documentation, not tested by this site.

Overview

The OpenFlowKit MCP server provides local-first diagramming tools for MCP clients like Claude, Cursor, Windsurf, and others. It offers deterministic validation, codebase analysis, templates, icon lookup, and viewer links without requiring an API key or cloud round-trip. Everything runs locally on the user's machine via a pure Node package distributed through npm.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →

NPM Sentinel MCP Server 79 · B

AI-powered NPM package analysis MCP server

★ 18 · Tools 19 Compare with this →

MartinLoop MCP Server 79 · B

An independent control layer that gives AI coding agents budgets, brakes, and receipts

★ 410 Compare with this →

GitHub MCP Server 76 · B

GitHub's official MCP server for managing repos, issues, PRs, and workflows via natural language

★ 33.5k · Tools 15 Compare with this →

Compare with other MCP servers

Source revision c9b5c77f2058 Data synced 2026-10-12 Read the FMRS scoring method