T1ENS3Ground Your Wukong — Model + KnowledgeWhat happens when you hit Send: YOU → the Agent → Wukong (the LLM — a brain in a jar: brilliant, no hands, no memory) · The Agent gathers three things every turn — persona card + chat history + your new message — into ONE stack: the context. "Memory" lives in the stack, not in Wukong · Token economics: every sheet in the stack is re-sent, re-read, re-billed every turn; the context window is the paper tray · The model menu: Claude / GPT / Gemini in three sizes each, and the open wave (DeepSeek #1 by token share) up to 20× cheaper — WukongDojo lets you pick any good LLM, your Wukong stays yours · You + Wukong = Your Wukong: your domain + your script beats generic ChatGPT on three axes — narrow · deep · real · Lab · Ground Your Wukong: attach one knowledge .md and watch the same model stop guessing★★★ Freedom