Bring MockFlow to your AI workflow with MCP

MCP (Model Context Protocol) is an open standard created by Anthropic that lets AI assistants like Claude, Cursor, and VS Code Copilot connect to external tools. Think of it as a universal plug that lets your AI talk to other software. With MockFlow's MCP servers, your AI can create flowcharts, wireframes, mind maps, and more — without you having to draw anything.

Mida AI is MockFlow's built-in AI assistant that lives inside the IdeaBoard and WireframePro editors. It generates, modifies, and transforms content directly on your canvas with full context of your project.

MCP servers let you use MockFlow's visualization tools from external AI tools like Claude Code, Cursor, or VS Code Copilot. MCP is ideal when you're already working in your IDE or AI chat and want to create visuals without switching to MockFlow.

In short: Mida AI is the full experience inside MockFlow. MCP brings MockFlow's tools to wherever your AI already is.

Loginless Web MCP — If you use Claude.ai or ChatGPT in the browser and just want quick visualizations. No installation, no login. Boards are created as shareable links.

Desktop App MCP — If you use the MockFlow Desktop App daily. Most powerful option — AI creates visuals directly on your open canvas. Includes exclusive tools like layout_board and read_wireframe.

CLI Web MCP — If you're a developer using Claude Code, Cursor, or VS Code Copilot. Boards are saved to your MockFlow account with OAuth login.

Loginless Web MCP: No account needed. Boards are created as anonymous shareable links.

Desktop App MCP: Yes, you need to be logged into the MockFlow Desktop App.

CLI Web MCP: Yes, you log in via browser (OAuth) when you run mockflow-ideaboard-mcp login. Boards are saved to your account.

IdeaBoard is MockFlow's visual brainstorming tool — an infinite canvas for flowcharts, mind maps, kanban boards, architecture diagrams, customer journey maps, storyboards, calendars, charts, and more. It has 16 MCP tools.

WireframePro is MockFlow's design tool for wireframing and UI design. Its MCP server can convert AI-generated HTML into editable wireframe components, and export wireframes back to AI for code generation. It has 3 MCP tools plus the read_wireframe export tool.

MockFlow MCP works with any MCP-compatible AI tool, including:

• Claude.ai (browser connector)
• Claude Code (terminal)
• Cursor (IDE)
• VS Code Copilot
• ChatGPT (via connector)
• Codex (OpenAI)
• Any tool that supports the Model Context Protocol

The Loginless Web MCP is completely free — no account. The CLI and Desktop App MCP use your MockFlow account, which has a free tier. The MCP npm packages are open source (MIT license) and free to use.

AI Skills are pre-built prompts that teach Claude Code how to use MockFlow MCP tools for specific tasks. For example, the "decision flowchart" skill automatically creates a yes/no decision tree when you ask "Should I use React or Vue?" Skills are available in the GitHub repo and the MockFlow Recipe Library.

Yes. The Desktop App MCP works with both cloud projects and local file-mode projects (.board and .wire files). Components are rendered directly on the canvas and saved to the file automatically.

Yes. Rate limits depend on how you connect and your plan. Only visualization tool calls are counted — protocol messages (initialize, tools/list, ping) are unlimited.

IdeaBoard:
• Loginless Web MCP: 5 visualizations per 30 minutes + 20 per day (per session)
• CLI MCP (Basic): 10 tool calls per 30 minutes (per user)
• CLI MCP (Paid): 30 tool calls per minute (per user)
• Desktop App (Basic): 10 tool calls per 30 minutes
• Desktop App (Paid / App Store): 30 tool calls per minute

WireframePro:
• CLI MCP (Basic): 10 tool calls per 30 minutes (per user)
• CLI MCP (Paid): 30 tool calls per minute (per user)
• Desktop App (Basic): 10 tool calls per 30 minutes
• Desktop App (Paid / App Store): 30 tool calls per minute