Coding setup?

[ltguillaume]

If I may ask: what is your software setup for developing this application? Which IDE/editor? I'm pretty bad at guessing (if and) how much use is made of AI, but I see some verbose issue replies, comments and docs that all do seem to suggest it. So if true, which model(s) do you use and what strategy do you adopt?

[dongdongbh]

I develop Mindwtr on Arch Linux using Neovim. I do not use a traditional IDE; my workflow is almost entirely open-source and CLI-based.

AI is part of the development process, but it does not replace engineering judgment or project ownership. The project’s stance is explained here: https://docs.mindwtr.app/start/faq#is-mindwtr-ai-generated-what-is-the-project-s-stance-on-ai

I generally use the most capable frontier models available; at present, that means Claude Fable 5 and GPT-5.6 Sol. These are paid services and represent a real ongoing development cost. I use them for tasks including design discussion, implementation, review, documentation, and debugging. I have also built and maintain project-specific workflows, rules, and skills so that the models work consistently with Mindwtr’s architecture and principles.

Using AI effectively still requires substantial review, testing, decision-making, and maintenance. It helps me move faster, but I continue to invest a great deal of time and effort into every part of Mindwtr.

[ltguillaume]

Thanks for giving a bit more insight in the development process, much appreciated!

I feel the workflow in terms of bug reports/feature requests is extremely quick, communicative and pleasant for end users, and it is obvious you have a clear scope for the application. Considering most projects I try to help out with a bit, this is something I do not take for granted. Just wanted you to know 🙂