MockFlow IdeaBoard MCP Server

Local MCP server for MockFlow IdeaBoard. Create flowcharts, mind maps, kanban boards, cloud architecture diagrams, and 12+ other visualization types — directly from AI-powered coding tools.

Works with Claude Code, Cursor, VS Code Copilot, Codex, and any MCP-compatible client.

Note: MCP servers may not provide the full experience for certain types of generation involving images, AI art, or complex designs. For the best seamless experience, use the Mida AI agent inside the MockFlow editor.

Quick Start

1. Install

npm install -g @mockflow/ideaboard-mcp

Or run without installing:

npx @mockflow/ideaboard-mcp

2. Authenticate

mockflow-ideaboard-mcp login

This opens your browser to MockFlow's login page. Log in with your MockFlow account and authorize access. The token is saved automatically to ~/.mockflow/credentials.json (one-time setup).

3. Start the Server

mockflow-ideaboard-mcp

You'll see:

MockFlow IdeaBoard - Local MCP Server
======================================
User: you@example.com

MCP server running on http://localhost:21193/mcp

Add to your AI client:

  Claude Code:
    claude mcp add --transport http -s user mockflow-ideaboard http://localhost:21193/mcp

4. Connect Your AI Client

Claude Code

claude mcp add --transport http -s user mockflow-ideaboard http://localhost:21193/mcp

Cursor

Settings > Cursor Settings > Tools & MCP:

{
  "mcpServers": {
    "mockflow-ideaboard": {
      "url": "http://localhost:21193/mcp"
    }
  }
}

VS Code Copilot

Create .vscode/mcp.json in your project:

{
  "servers": {
    "mockflow-ideaboard": {
      "type": "http",
      "url": "http://localhost:21193/mcp"
    }
  }
}

Codex (OpenAI)

codex mcp add mockflow-ideaboard http://localhost:21193/mcp

5. Start Prompting

Ask your AI client:

"Create a flowchart showing the user registration process"
"Create a kanban board for the product launch"
"Draw an AWS architecture diagram with API Gateway, Lambda, and DynamoDB"
"Create a mind map about project management methodologies"
"Map out the database schema as an ER diagram"

The server creates a new IdeaBoard and returns the URL. Open it in your browser to view and edit.

Available Tools (16)

How It Works

You: "Create a kanban board for the sprint"
  |
  v
AI Client (Claude Code / Cursor / VS Code)
  |  generates structured data (columns, cards, etc.)
  v
MCP Server (localhost:21193)
  |  proxies to MockFlow backend
  v
app.mockflow.com
  |  creates IdeaBoard with visualization
  v
Returns board URL → open in browser to view/edit

1. Your AI client reads your codebase or follows your prompt
2. It calls the appropriate render_* tool via MCP with structured data
3. The local MCP server forwards the request to app.mockflow.com
4. MockFlow creates the board and returns a URL
5. You open the URL to view, edit, and share the visualization

CLI Reference

mockflow-ideaboard-mcp                     # Start server on default port (21193)
mockflow-ideaboard-mcp --port=8888         # Start on custom port
mockflow-ideaboard-mcp login               # Set up API key (one-time)
mockflow-ideaboard-mcp --help              # Show usage and setup instructions

Configuration

Credentials

Stored in ~/.mockflow/credentials.json (created automatically by mockflow-ideaboard-mcp login):

{
  "access_token": "...",
  "userid": "you@example.com",
  "clientid": "your-client-id"
}

Custom Port

If port 21193 is in use:

mockflow-ideaboard-mcp --port=8888

Then update your AI client config to use the new port.

Debug Mode

For verbose logging:

MCP_DEBUG=1 mockflow-ideaboard-mcp

Verify Installation

# Check server is running
curl http://localhost:21193/mcp

# List available tools
curl -X POST http://localhost:21193/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list","params":{}}'

Rate Limits

Only visualization tool calls (e.g. render_flowchart, render_mindmap) are counted. Protocol messages (initialize, tools/list, ping) are unlimited.

Limits are enforced server-side per authenticated user. Every response includes RateLimit-Limit, RateLimit-Remaining, and RateLimit-Reset headers.

Troubleshooting

"No credentials found"

Run mockflow-ideaboard-mcp login to authenticate with your MockFlow account.

Port already in use

Another process is using port 21193. Either:

• Use a different port: mockflow-ideaboard-mcp --port=8888
• Kill the process: lsof -ti :21193 | xargs kill

AI client doesn't use IdeaBoard tools

Be explicit in your prompt: "Using mockflow-ideaboard, create a flowchart for...". AI clients have many tools and may default to generating code.

Tool call fails with backend error

Ensure you have a valid login and internet connection. The server needs to reach app.mockflow.com. If your token expired, run mockflow-ideaboard-mcp login again.

Example Prompts

Codebase Visualization

"Analyze this codebase and create a system architecture diagram"
"Create a flowchart showing the authentication flow in this project"
"Map out the database schema as an ER diagram"
"Visualize the API endpoints as a swimlane diagram"

Project Management

"Create a kanban board for the remaining TODOs in this project"
"Create a gantt chart for the Q2 product roadmap"
"Create a calendar with the team's sprint events for April"

Brainstorming

"Create a mind map about microservices architecture patterns"
"Create a whiteboard with sticky notes for the retrospective"
"Create a customer journey map for the onboarding experience"

AI Skills (Claude Code)

Skills are pre-built prompts that teach Claude Code how to use IdeaBoard MCP tools for specific tasks. Instead of writing detailed prompts every time, skills trigger automatically when Claude detects a matching intent.

Install a Skill

Copy a skill file to your Claude Code skills directory:

mkdir -p ~/.claude/skills/mockflow-decision-flowchart
# Copy SKILL.md into the folder (see skills/ directory in this repo)

Available Skills

More skills available in the MockFlow Recipe Library — each recipe can be used as a Claude Code skill.

Create Custom Skills

Use the AI Prompt Box inside IdeaBoard to create and test custom prompts, then export them as Claude Code skills. See the skills/ directory for the SKILL.md format.

Skill Format

---
name: skill-name
description: When to use this skill
allowed-tools: MCP(render_flowchart)
user-invocable: true
---

# Instructions for Claude Code
...

Links

• npm: npmjs.com/package/@mockflow/ideaboard-mcp
• GitHub: github.com/mockflow/ideaboard-mcp
• MockFlow: mockflow.com
• WireframePro MCP: @mockflow/wireframepro-mcp — convert HTML to wireframes

Contributing

Issues and pull requests are welcome at github.com/mockflow/ideaboard-mcp.

License

MIT