MCP (connector)

[ashrafabulsoud]

MCP Connector for Generating and Embedding Diagrams in AI Outputs

Idea

I would like to suggest adding an MCP connector for OpenFlowKit, so AI tools can generate diagrams directly inside OpenFlowKit while creating plans, PRDs, technical documents, workflows, architecture documents, or project roadmaps.

Problem

When using AI to generate product plans, PRDs, technical architecture, or process documentation, diagrams are often needed to explain the idea clearly.

Currently, the AI can describe the diagram in text, but the user usually has to manually recreate it in a diagramming tool later. This creates extra work and makes the generated documentation less complete.

Proposed Solution

Create an OpenFlowKit MCP connector that allows AI assistants and coding agents to:

1. Generate diagrams directly in OpenFlowKit.
2. Save the generated diagram inside the user's OpenFlowKit workspace.
3. Return a link to the generated diagram.
4. Embed that link inside the generated PRD, plan, report, or documentation file.

For example, when the AI generates a PRD, it could automatically include:

• User flow diagram
• System architecture diagram
• Database relationship diagram
• Process/workflow diagram
• Feature flow diagram
• Integration diagram

And then include a reference like:

## System Architecture

Diagram: [Open in OpenFlowKit](generated-diagram-link)

Example Use Case

A user asks an AI assistant:

Create a PRD for a booking system with user roles, payment flow, and admin approval workflow.

The AI would generate the PRD, and also call the OpenFlowKit MCP connector to create diagrams such as:

• Booking workflow
• Payment flow
• Admin approval process
• High-level system architecture

Then the final PRD would contain links to the generated diagrams inside OpenFlowKit.

Why This Is Useful

This would make OpenFlowKit more powerful as part of AI-assisted planning and documentation workflows.

Instead of producing text-only documentation, AI tools could generate complete, visual, and actionable documents.

This is especially useful for:

• Product managers
• CTOs
• Developers
• System architects
• Business analysts
• QA teams
• Startup founders
• AI coding agents

Suggested MCP Capabilities

The connector could support actions like:

create_diagram
update_diagram
get_diagram_link
export_diagram
list_diagrams

Possible input formats:

• Mermaid
• JSON schema
• Plain-text diagram description
• Flow definition
• Nodes and edges structure

Possible output:

{
  "diagram_id": "123",
  "diagram_url": "https://openflowkit.com/...",
  "embed_url": "https://openflowkit.com/embed/..."
}

Expected Benefit

This would allow OpenFlowKit to become a natural diagram-generation layer for AI tools.

The AI would not only explain ideas, but also create usable diagrams that remain editable and shareable inside OpenFlowKit.

I believe this could be a strong use case for OpenFlowKit, especially as AI-assisted PRD generation, planning, and technical documentation become more common.

[Vrun-design]

Thanks for suggesting this. I think an OpenFlowKit MCP connector makes a lot of sense, especially as MCP is becoming a standard way for AI tools to connect with external apps and workflows.

We’re also exploring the new web MCP direction and how that could fit into OpenFlowKit. Stay tuned — we may have something to share in the coming weeks.