Using BMAD workflow without MAX models in cursor

[mzafarr]

Is it possible to use entire BMAD workflow effectively without issues with normal models and not MAX models in cursor. Are there adjustments to do for MAX and regular models? Correct me I am wrong, I think with all these docs and all maybe its best suited for MAX models only. Need tips for users like me who don't use MAX models. Thank you.

[bmadcode]

@mzafarr I do not use the max models and have had great success - but I do:

• Do all planning and analysis through PRD and architectures in the Web
• Ensure that I generate the story with the dedicate sm.ide.md
• develop with the dev.ide.sm

I am working on some optimizations that I hope will further improve the context overhead on the dev agent, as it does fill up on context in sonnet 4 non max relatively quickly so have to keep the stories and tasks small. Hope the changes I am thinking about will help. But. still, following now as is I have luck not needing to use max mode.

[mzafarr]

Wonderful by the way. Is it okay if we create stories within the web agent (the one in web-build-sample folder)? I have been doing it with Gemini in the Web, and I thought that I might get more context if I do it in the Web, which might produce better stories. I'm not 100% sure. Is that so?

[bmadcode]

Ive not experimented much with doing the detailed stories in the web as it just seems to be more of a pain, mainly because I work in a tight loop of draft story, dev story, repeat until done. and the sm.ide and dev.id are good at finding the files they need to produce and dev from the stories 1 at a time.

I have not fully optimized generating stories in the web - especially more than 1 at a time.

not to say it wont work, I am just not sure its going to be a great experience. Also also once getting into dev, you are going to run into something or have an idea that is going to then require major or minor changes to the rest of the stories.

One potential way to possibly make the web better for producing stories, is still doing 1 at a time, but before drafting the next story, you upload or paste the copy of the completed story back so it sees and change notes or issues from the dev on the recently created story