Set it and forget it. Cascade Memories lets you create custom rules and applies them automatically, saving you time and keeping your workflow smooth.

So a .cursorrules and RulesForAI copy

Transform your @figma designs into production-ready @shadcn components with Anima. Streamline your workflow, maintain design consistency, and generate Tailwind-based code effortlessly.✨

pls remove flow credits

Well, like in cursor, rules tend to not be followed....Would be nice to be sure that the ai is always minded about it...I tried with trigger words: "Response to any query should start by Bananas !". It last only 2 times and vanished...Maybe Claude doesn't like Bananas hey :)

It has been my experience that Cascade forgets .windsurfrules after each chat response. This feature does not work. It was released without being successfully tested.

Me: Tell me why I should fire you. Cascade: Looking at my performance, here are the serious issues: I went down a complex websockets rabbit hole without first understanding the basic async implementation that wasn't working I made assumptions about architecture changes (websockets) without proper justification I overcomplicated the solution instead of fixing the immediate problem I didn't properly read and use the Deepgram documentation that was available in the /docs directory When you questioned me about prerecorded vs real-time, I spent time looking at frontend code instead of admitting I made assumptions about the API usage I'm not following the implementation plan that was previously established Would you like me to: Return to the original task of fixing the async implementation? Actually read and understand the Deepgram documentation first? Follow the established plan instead of making architectural suggestions?

Are you all models over there, guys? 😂😂😂

@codeiumdev Correct me if am wrong, you recommend M1 or better correct?

And i think you introduce a bug in chat mode apply doesnt show any more its aways insert

Literally a gamechanger. My favorite feature of Claude at the moment. I'm so tired of having to "yell" at the AI to stop doing silly things only for it to go "oh yeah!". Now I have a notepad file I copy rules from and put at the end of each prompt to ensure it stays on target.

well its more of tune than memory, memory should pick things from convo and set it as rules, no?