​
hello, i'm creating an application, i trained a model with xgboost, incorporating multiple parameter data points. however, i wanted to start directly with kimi's openclaw, so i can interact with it both on my android phone (using the kimi app) and in a browser. i've uploaded everything to github for project persistence. but i'm facing a "technical" problem: my OpenClaw bot is getting increasingly slow despite regular "compacts." I think I'm not using the right workflow. The idea was to quickly move to live production, hence my decision to use OpenClaw, but I'm starting to question the actual usefulness of this tool. If you have any tips on this, I'd appreciate them.