Integration with MCP (model-context-protocol) for LLMs such as Claude

[ansis-m]

• I agree to follow the Code of Conduct that this project adheres to.
• I have searched the issue tracker for a feature request that matches the one I want to file, without success.

Is integration with MCPs in progress? If not, is it in planned in the future (either for web or desktop)? I think Claude would be totally capable to iterate and improve when provided the output in the svg or jpeg/png format.

I know that there are many programmatic approaches for generating draw.io files so MCP integration seems like a logical next step.

I think this would be a great hobby-project which I would be willing to do or at least contribute. However, I probably would not want to spend time on project that's already in progress or finished by someone else.

Thank you in advance and excuse me if I wrote something stupid!

[rouxianmantou]

Any progress now?
I am also thinking about this, but I just started thinking about it today and asked Claude what I should do and where the difficulty is... 😂😂😂

[alderg]

That? #4044